WO2017203681A1

WO2017203681A1 - Information processing device, information processing method, program, and storage medium

Info

Publication number: WO2017203681A1
Application number: PCT/JP2016/065693
Authority: WO
Inventors: 史和藤原
Original assignee: 楽天株式会社
Priority date: 2016-05-27
Filing date: 2016-05-27
Publication date: 2017-11-30
Also published as: JP6539782B2; JPWO2017203681A1

Abstract

The purpose of the present invention is to prevent discrepancy between the content of an evaluation sentence and an evaluation value on the premise that a user inputs the evaluation value himself, so as to improve reliability of evaluation information. In the present invention, a sentence inputted by a user is put to analysis, and an upper limit and a low limit are set for an evaluation value on the basis of the number of high-evaluation words and low-evaluation words included in the input sentence, and when an input of the evaluation value from the user is accepted, an input permission range for the evaluation value is limited to within a range between the lower limit and the upper limit thus set, and at the same time, information prompting revisions to be made to the input sentence is presented in response to an input of an operation for instructing input of an evaluation value falling outside the input permission range, and an input sentence for which an operation to reject revision has been accepted is analyzed, and then a revision is made to criteria for setting the upper limit and the lower limit.

Description

Information processing apparatus, information processing method, program, and storage medium

The present invention relates to an information processing apparatus and method, a program for realizing the information processing apparatus, and a storage medium storing the program, and in particular, a technical field for improving the reliability of evaluation information on products and the like. About.

JP 2014-21578 A

As so-called online shopping, for example, it is possible to purchase products at a shopping site constructed on the Internet.
Since online shopping is not possible to pick up the actual product, the shopping site provides review information (evaluation information) on the product posting page, etc., and assists in selecting the purchased product. Yes. As the review information, for example, a user who has actually purchased a product can input an impression about the product (for example, an impression about a feeling of use) as a sentence, or can input an evaluation value based on the number of stars.

Since the evaluation information based on these sentences and values functions as important reference information when the user as a general consumer selects a purchased product, it is required to be input accurately.
In practice, however, there are some cases in which there is a discrepancy between the text content and the evaluation value, such as a low evaluation value even though the text content is highly evaluated. Is done.

Patent Document 1 discloses a technique for calculating an evaluation value obtained by analyzing an evaluation comment input by a user by natural language processing and evaluating a comment content, and automatically setting the evaluation value for the comment. Yes.

However, the technique described in the above-mentioned Patent Document 1 automatically sets an evaluation value, and does not prevent a discrepancy between the content of the evaluation text and the evaluation value when the user himself inputs the evaluation value.

The present invention has been made in view of the above circumstances, and on the premise that the user himself inputs the evaluation value, the content of the evaluation text and the evaluation value are prevented from being flawed, and the reliability of the evaluation information is improved. The purpose is to plan.

An information processing apparatus according to the present invention analyzes an input sentence by a user, and sets an upper limit value and a lower limit value of an evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence. When the input of the evaluation value from the user is accepted, the input permission range of the evaluation value is limited to the range from the lower limit value to the upper limit value set by the upper and lower limit value setting unit, and the input permission In response to an operation input for instructing input of the evaluation value outside the range, an evaluation input restriction unit for presenting information for prompting correction of the input sentence, a correction rejection receiving unit for receiving the correction rejection operation, and the correction A reference learning unit that analyzes an input sentence in which the rejection operation is received by a rejection reception unit, and corrects the upper limit value and the lower limit value setting criterion by the upper and lower limit value setting unit;

According to the said structure, the range of the evaluation value in which the input by a user is permitted is restrict | limited according to the content of an input sentence. That is, it is possible to prevent mistakes in evaluation value input while leaving room for the user to set evaluation values.
Moreover, the setting accuracy of the upper limit value and the lower limit value can be improved by the learning function based on the analysis of the input sentence whose correction is rejected.

In the information processing apparatus according to the present invention described above, the reference learning unit can correct the setting criterion according to the number of rejections for input sentences including the same high evaluation word or low evaluation word. .

This prevents the upper and lower limit setting criteria from fluctuating due to noise when only one user rejects the correction.

In the information processing apparatus according to the present invention described above, the reference learning unit corrects the setting criterion by a plurality of users having different attributes to reject input sentences including the same high evaluation word or low evaluation word. Depending on the situation.

This makes it possible to correct the setting standard in response to the possibility that the setting standard is more likely to be incorrect.

In the information processing apparatus according to the present invention described above, the upper / lower limit value setting unit may mean either a high evaluation word or a low evaluation word based on a nearby context for a high / low evaluation determination difficult word detected from the input sentence. Can be used for setting the upper limit value and the lower limit value as a word meaning the estimated one.

This makes it possible to increase the number of high evaluation word and low evaluation word samples required for setting the upper and lower limit values.

In the information processing apparatus according to the present invention described above, the evaluation input restriction unit can present information for specifying the high evaluation word and the low evaluation word in the input sentence to the user.

This makes it possible to suggest to the user the part that became the basis for setting the upper and lower limits in the input text.

In the information processing apparatus according to the present invention described above, the evaluation input restriction unit is designated by the input operation when an input operation for designating a value outside the input permission range is performed as the evaluation value input operation. It is possible to present to the user information indicating a point requiring correction in the input sentence so as to make it possible to input an evaluation value with the set value.

This makes it possible to suggest to the user which part of the input sentence should be corrected in order to allow the input of the desired evaluation value.

In the information processing apparatus according to the present invention described above, the evaluation input restriction unit is designated by the input operation when an input operation for designating a value outside the input permission range is performed as the evaluation value input operation. It is possible to present to the user another input sentence in which the evaluation value input with the same value as that of the input value is permitted.

This makes it possible to suggest to the user what text correction should be performed in order to allow the input of the desired evaluation value.

In the information processing apparatus according to the present invention described above, the input text and the evaluation value input from the user are received for each item in a plurality of evaluation items, and separately from the evaluation value for each item. The upper and lower limit value setting unit accepts an input of a comprehensive evaluation value, and the upper and lower limit value setting unit sets an upper limit value and a lower limit value for the comprehensive evaluation value according to the number of high evaluation words and low evaluation words in the entire input sentence over the plurality of items. The evaluation input restriction unit sets the input permission range for the comprehensive evaluation value and presents information for urging the correction of the input sentence, and sets the input sentence and the input evaluation value for each item. It is possible to further include a wrinkle item notifying unit that determines whether or not there is a flaw and notifies the user of the item determined to have a flaw.

When the evaluation for each item and the comprehensive evaluation are possible, it is estimated that the evaluation value and the text content are the same as the overall evaluation. There may also be cases in which wrinkles occur.
According to the above configuration, it is possible to notify the user of the occurrence of wrinkles and their locations on an item basis corresponding to such a case.

Further, the information processing method according to the present invention analyzes an input sentence by a user, and sets an upper limit value and a lower limit value of an evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence. When accepting an input of the evaluation value from the setting step and the user, the input permission range of the evaluation value is limited to a range from the lower limit value to the upper limit value set by the upper and lower limit value setting step, and In response to an operation input for instructing input of the evaluation value outside the input permission range, an evaluation input restriction step for presenting information for prompting correction of the input sentence, a correction rejection reception step for receiving the correction rejection operation, The input sentence in which the rejection operation is accepted in the correction rejection acceptance step is analyzed, and the setting criteria for the upper limit value and the lower limit value in the upper / lower limit value setting step are corrected. The reference learning step for an information processing method by the information processing device executes.

Also by the information processing method according to the present invention, the same operation as the information processing apparatus according to the present invention described above can be obtained.

Furthermore, a program according to the present invention is a program that causes an information processing apparatus to execute processing executed as the information processing method.
Furthermore, a storage medium according to the present invention is a storage medium storing the program.
The above information processing apparatus is realized by these programs and storage media.

According to the present invention, it is possible to prevent the discrepancy between the content of the evaluation text and the evaluation value under the premise that the user himself inputs the evaluation value, and to improve the reliability of the evaluation information.

It is the figure which showed the example of the network system of embodiment. It is a block diagram of the computer apparatus of an embodiment. It is the functional block diagram which blocked and showed the function concerning the review contribution which the information processor of an embodiment has. It is the figure which showed the example of the web page for review contribution. It is the figure which illustrated the mode of the web page in the state where the input of the evaluation text with respect to the evaluation value input column was performed. It is the figure which illustrated the storing information of evaluation word DB. It is the figure which illustrated the storage information of difficult-to-determine word DB. It is the figure which showed the example of the web page used as the state into which evaluation value input is possible. It is the figure which showed the example of the web page when evaluation value input within the input permission range was performed. It is the figure which illustrated the storage information of review DB. It is the figure which showed the example of the alert display for prompting correction of an input sentence. It is the figure which illustrated the display content of the web page corresponding at the time of sentence correction. It is the figure which showed the example of the reference display. It is the figure which illustrated the display content of the web page according to correction refusal operation. It is the figure which illustrated storage information (rejection review management information) of rejection review DB. It is the figure which illustrated word refusal count management information. It is the flowchart which showed the procedure of the process from the input of review information to registration of review information. Similarly, it is a flowchart showing a processing procedure from input of review information to registration of review information. It is the flowchart which showed the process sequence for learning of the setting reference | standard of an upper limit and a lower limit. It is the figure which illustrated the storage information (rejection review management information) of rejection review DB in the modification concerning the correction trigger of adjustment value. It is the flowchart which showed the procedure of the process which should be performed in the modification concerning the correction trigger of an adjustment value. It is the figure which showed the example of the web page in 2nd Embodiment. It is an example of a state in which text input and evaluation value input for each evaluation item are performed on the web page in the initial state. It is the figure which showed the example of the display content of the web page shown according to the item evaluation completion button being operated. It is the figure which illustrated the mode of the web page when operation which instruct | indicates the evaluation value input within the input permission range about comprehensive evaluation value was performed. It is the figure which showed the example of the wrinkle in the item unit. It is the figure which showed the example of the 齟齬 alert display. It is the flowchart which showed the procedure of the process which should be performed in order to implement | achieve the function as 2nd Embodiment.

Hereinafter, embodiments will be described in the following order.

<1. First Embodiment>
[1-1. Network system]
[1-2. Hardware configuration of computer device]
[1-3. Function related to review posting]
[1-4. Processing procedure]
[1-5. Modifications related to adjustment value correction trigger]
[1-6. Summary of First Embodiment]
<2. Second Embodiment>
[2-1. Function and processing procedure]
[2-2. Summary of Second Embodiment]
<3. Program and Storage Medium>
<4. Modification>

<1. First Embodiment>
[1-1. Network system]

FIG. 1 shows an example of a network system 1 assumed in the present embodiment. The network system 1 functions as an EC (EC: electronic commerce) system.
The network system 1 is configured such that a service providing system 3 and a plurality of

user terminals

4, 4,... Can communicate with each other via a network 2 as the Internet, for example.

The user terminal 4 is a computer device provided with a web browser. Examples of the user terminal 4 include a high-function mobile phone (smart phone), a mobile phone, a personal digital assistant (PDA), and a portable or stationary personal computer (PC). It is not limited.
The user terminal 4 transmits an HTTP (Hypertext Transfer Protocol) request to an external server device capable of communicating via the network 2, particularly, in this example, to the service providing server 30 in the service providing system 3. Request a predetermined process. The user terminal 4 receives the web page sent in response to the HTTP request and displays it on the web browser. Thereby, the user can browse and operate a desired web page.

The service providing system 3 includes a service providing server 30, an evaluation word DB (database) 31, a difficult-to-determine word DB 32, a review DB 33, and a rejection review DB 34 each constituted by a computer device. Each of these devices can communicate with each other via a network such as a LAN (Local Area Network).

The service providing server 30 performs various processes based on the HTTP request sent from the user terminal 4 that can communicate with the network 2. For example, generation and transmission of various web pages (for example, a product web page, an order web page, a review posting page, etc.), a purchase process according to an order confirmation operation by the user, and the like are executed.

In the network system 1, a website related to product sales is provided to the user (user of the user terminal 4) by the service providing server 30. Here, in the present specification, the product is a concept including not only tangible objects but also intangible objects such as electronic data such as music and videos, and various services such as accommodation services for travel and business trips.
Hereinafter, a website related to product sales provided by the service providing server 30 is also referred to as a “service providing site” in the sense of a website for providing a product sales service to a user.

A plurality of stores (product sales stores) or facilities (facility that provides services such as accommodation services) are members of the service providing site. The staff of each store or facility registers the product of the own store or the own facility in the service providing system 3 via the computer device, so that various products from various stores or facilities are uploaded on the EC site. The user can purchase a desired product by accessing the service providing site through the user terminal 4.

Here, in this example, the user can perform membership registration for the service providing system 3 when using the service providing site. At the time of member registration, the user registers necessary information such as a user ID (user identification information), merchandise destination information (user address information), and a credit card number. The user logs in to the service providing site with the registered user ID, thereby saving the trouble of inputting necessary information again when purchasing a product on the service providing site.

In response to a request from the user terminal 4, the service providing server 30 transmits a web page (hereinafter referred to as “review posting page”) for posting a review about the product.
In this example, a review posting page is prepared for each product, and the service providing server 30 transmits a review posting page for the product instructed by the user terminal 4. Details of the review posting page will be described later.

Here, a predetermined web page in the service providing site is provided with a button (URL link) for requesting a review posting page. Specifically, in this example, the button is provided for each product on a purchase history page that lists a list of products purchased by the user in the past. The user clicks a button provided on the purchase history page corresponding to the product to be posted for review to transmit a review posting page for a desired product among the purchased products. Can request (instruct) (page request).
In the case of this example, a button for requesting a review posting page is also provided on each product web page, and the user accesses the desired product web page and clicks on the button for the desired product. It is also possible to request (instruct) to send a review posting page.
In the service providing system 3 of this example, when a user posts a review, it is assumed that the user logs in with his / her user ID.

Review information posted on the review posting page is registered (stored) in the review DB 33 by the service providing server 30 and managed.
Details of the review DB 33 will be described later.

The review information stored in the review DB 33 is presented in response to a request from the user terminal 4.
Specifically, in this example, the product web page on which the product is posted is provided with a button for browsing the review information. The service providing server 30 provides HTML (HyperText Markup Language) data for presenting a review browsing page for browsing the review information posted on the corresponding product in response to the button being operated by clicking or tapping. And transmit the data to the corresponding user terminal 4.

The evaluation word DB 31 stores a list of evaluation words. The evaluation word is a word estimated to be used to express high / low evaluation among words that are supposed to be written as a review. For example, words such as “good”, “best”, “worst”, and “bad” are applicable. Details of the information stored in the evaluation word DB 31 will be described later.

The difficult-to-determine word DB 32 stores a list of high / low evaluation difficult-to-determine words, which will be described later. Further, the rejection review DB 34 stores information related to the review posted after the correction of the text content is rejected as will be described later.
The details of the information stored in the difficult-to-determine word DB 32 and the rejection review DB 34 will be described again.

In FIG. 1, various examples of the configuration of the network 2 are assumed. For example, the Internet, intranet, extranet, LAN (Local Area Network), CATV (Community Antenna TeleVision) communication network, virtual private network, telephone line network, mobile communication network, satellite communication network, etc. Is assumed.
Various examples of transmission media constituting all or part of the network 2 are also envisaged. For example, IEEE (Institute of Electrical and Electronics Engineers) 1394, USB (Universal Serial Bus), power line carrier, telephone line, etc. can be wired with infrared, IrDA (Infrared Data Association), Bluetooth (registered trademark), 802.11 wireless It can also be used by radio such as a mobile phone network, satellite line, and digital terrestrial network.

In addition, the service providing server 30 in the service providing system 3 does not need to be configured by a single computer device, but may be configured by two or more computer devices.
In FIG. 1, the evaluation word DB 31, the difficult-to-determine word DB 32, the review DB 33, and the rejection review DB 34 are shown as separate devices for convenience of illustration and explanation. However, these DBs can be accessed by the service providing server 30. Any form may be used.

[1-2. Hardware configuration of computer device]

FIG. 2 shows a hardware configuration of a computer device that constitutes each device (evaluation word DB 31, determination difficult word DB 32, review DB 33, rejection review DB 34, and user terminal 4) including the service providing server 30 shown in FIG. Indicates.
In FIG. 2, a CPU (Central Processing Unit) 101 of a computer apparatus performs various processes according to a program stored in a ROM (Read Only Memory) 102 or a program loaded from a storage unit 108 to a RAM (Random Access Memory) 103. Execute the process. The RAM 103 also appropriately stores data necessary for the CPU 101 to execute various processes.
The CPU 101, ROM 102, and RAM 103 are connected to each other via a bus 104. An input / output interface 105 is also connected to the bus 104.
The input / output interface 105 includes an input device 106 such as a keyboard, a mouse, and a touch panel, a display (display device) such as an LCD (Liquid Crystal Display), a CRT (Cathode Ray Tube), an organic EL (Electroluminescence) panel, and a speaker. Are connected to a storage unit 108 configured by an HDD (Hard Disk Drive) or a flash memory device, and a communication unit 109 for mutual communication with an external device.
A media drive 110 is also connected to the input / output interface 105 as necessary, and a removable medium 111 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory is appropriately mounted, and information can be written to the removable medium 111. Reading is performed.

In such a computer apparatus, data and programs can be uploaded and downloaded through communication by the communication unit 109, and data and programs can be transferred via the removable medium 111.
When the CPU 101 performs processing operations based on various programs, particularly in the computer device as the service providing server 30, information processing and communication described below are executed.

[1-3. Function related to review posting]

Next, among the functions of the service providing server 30, various functions related to review posting will be described with reference to FIGS.
FIG. 3 is a functional block diagram showing the functions related to the review posting of the service providing server 30 in a block form.
As illustrated in FIG. 3, the service providing server 30 functions as an upper and lower limit setting processing unit F1, an evaluation input restriction processing unit F2, a correction rejection acceptance processing unit F3, and a reference learning processing unit F4 as functions related to review posting. Has each function.

The upper and lower limit value setting processing unit F1 analyzes an input sentence by the user and sets an upper limit value and a lower limit value of the evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence.

FIG. 4 shows an example of a web page wp for review posting in this example.
In the following explanation, an example in which the products sold on the service providing site are travel products as accommodation services provided by accommodation facilities such as hotels and inns, and the review targets are those accommodation facilities. To do.

As shown in FIG. 4, the web page wp in this case includes an evaluation value input field Rv for inputting an evaluation value, an evaluation text input field Rt for inputting an evaluation text, and the user completing the text input. An input completion button B1 for instructing is presented.
In the case of this example, the evaluation value is input by the number of stars as shown in the figure, and five stars are the highest (full score).
In the evaluation text input field Rv, a template for inputting an evaluation text for each evaluation item is adopted as a template. Specifically, in this example, in order to input the evaluation text individually for each evaluation item of “bath”, “meal”, “room”, “location”, and “service”, corresponding to the review for the accommodation facility The template is adopted.

The display content of the web page wp shown in FIG. 4 is the content presented to the user in response to the request for the review posting page described above. That is, the display content is the initial state when the review information is not input.
In the web page wp in the initial state, the evaluation value input to the evaluation value input field Rv is impossible. Specifically, in this example, the star in the evaluation value input field Rv is grayed out, and the evaluation value cannot be input.
As will be described later, in this example, the state transitions to a state in which an evaluation value can be input to the evaluation value input field Rv in response to the completion of the text input to the evaluation text input field Rt.

FIG. 5 illustrates a state of the web page wp in a state where the evaluation text is input to the evaluation value input field Rv.
The user operates the input completion button B1 when the input of the evaluation text in the evaluation value input field Rv is completed. Thereby, the display content of the web page wp can be changed to the display content according to the next state.

The upper and lower limit setting processing unit F1 shown in FIG. 3 analyzes the input sentence for the evaluation sentence input field Rt in response to the operation of the input completion button B1 as described above, and is included in the input sentence. The upper limit value and the lower limit value of the evaluation value are set based on the number of high evaluation words and low evaluation words.
At this time, when specifying the high evaluation word and the low evaluation word in the input sentence, the evaluation word list information stored in the evaluation word DB 31 is used.

FIG. 6 shows an example of information stored in the evaluation word DB 31.
In the evaluation word DB 31, an evaluation word selected in advance and word type information (high evaluation or low evaluation) of the evaluation word are associated with each other. Furthermore, each evaluation word is also associated with an adjustment value Va used when setting an upper limit value and a lower limit value of the evaluation value.

The upper and lower limit value setting processing unit F1 specifies a word that matches the evaluation word stored in the evaluation word DB 31 among the words in the input sentence. And the upper limit value and lower limit value of the evaluation value according to the content of the input sentence are set using the number of specified evaluation words and the adjustment value Va associated with the evaluation word.
Note that the evaluation word associated with “high evaluation” as the “word type” information is “high evaluation word”, and the evaluation word associated with “low evaluation” is “low evaluation word”.

In the case of this example, the upper limit value and the lower limit value are set to the total number of evaluation words (that is, the specified high evaluation word number + low evaluation word number) for the high evaluation word and the low evaluation word specified from the input sentence. It is performed according to the ratio (%) of the number of high evaluation words and the number of low evaluation words.
For example, “the ratio of the number of high evaluation words−the ratio of the number of low evaluation words” is calculated as the range setting reference value Vs, and the input permission range is set to “1 to 3” “2” according to the size of the range setting reference value Vs. ~ 4 "or" 3-5 ".
For example, if the range setting reference value Vs is less than −30 (%), the input permission range = “1 to 3” (that is, the lower limit value = 1, the upper limit value = 3), and the range setting reference value Vs = −30 to 30% If it is less, the input permission range = “2 to 4” (that is, the lower limit value = 2, the upper limit value = 4), and if the range setting reference value = 30 or more, the input permission range = “3 to 5” (that is, the lower limit value = 3 and upper limit value = 5).
In addition, the said numerical value is an example to the last, and is not limited to these numerical values.

In the method of setting the upper limit value and the lower limit value as described above, in this example, the adjustment value Va for each word is used as the ratio value of the number of high evaluation words / total number of evaluation words and the number of low evaluation words / total number of evaluation words. to add. Specifically, the adjustment value Va stored in the evaluation word DB 31 is added.
For example, the adjustment value Va is “high”, “very good”, “good”, which is a high evaluation word, “bad”, “very bad”, “worst”, etc., which are low evaluation words, and the adjustment value Va = “ “10” and “very good” are adjustment values Va = “5”, “good” is an adjustment value Va = “0”, “bad” is an adjustment value Va = “0”, and “very bad” is an adjustment value Va = “0”. For “−5” and “worst”, adjustment values Va = “− 10”, etc., and different values are set according to the meaning of the evaluation word (meaning related to the magnitude of evaluation high / low). As a result, an appropriate range setting reference value Vs can be calculated according to the meaning of the high evaluation word and the low evaluation word, and the input permission range can be set with higher accuracy.

Note that the range setting reference value Vs may be obtained as a difference between the detected high evaluation word number and the low evaluation word number, instead of using the above-described “ratio” difference. In this case, it goes without saying that a value corresponding to the change should be set as the adjustment value Va.

Also, in this example, when setting the input permission range (that is, the upper limit value and the lower limit value), it is considered that a high / low evaluation difficult word can exist in the input sentence. Here, the high / low evaluation difficult word is a word represented by, for example, “Yabai” and can function as an evaluation word, but the word alone means either high evaluation word / low evaluation word. It means a word that is difficult to determine whether it is used. In addition to “Yabai”, for example, words such as “Not a crap” and “Danger” can be cited. For languages other than Japanese, words such as “sick” (English) can be used.

Such a high / low evaluation difficult word is presumed to be used by the user to express high / low evaluation. Therefore, in setting an appropriate input permission range according to the text content, It should be used for setting the input permission range as a word corresponding to either the high evaluation word or the low evaluation word. In particular, in the case of assuming relatively short text input as in this example, the number of evaluation words detected in the text tends to be small, so in setting an appropriate input permission range, as much as possible It is desirable to be able to sample evaluation words.

For this reason, in this example, the high / low evaluation difficult word is detected for the input sentence, and the detected high / low evaluation difficult word is used in any meaning of the high evaluation word / low evaluation word. Is estimated from the context in the vicinity of the high / low evaluation difficult word, and is used for setting an upper limit value and a lower limit value as a word having the estimated meaning.

In the high / low evaluation difficult-to-determine word detection process, information stored in the difficult-to-determine word DB 32 as shown in FIG. 7 is used.
In FIG. 7, the hard-to-determine word DB 32 stores a list of high / low-evaluation difficult words, and a high-evaluated neighborhood word and a low-evaluated neighborhood word are associated with each high / low-evaluated difficult word. Stored.

When the high evaluation neighborhood word and the low evaluation neighborhood word are used in the vicinity of the high / low evaluation judgment difficult word, the meaning of the high / low evaluation judgment difficult word is the meaning of the high evaluation side, It is a word that is presumed to be meaningful.
For example, the high / low evaluation difficult word “Yabai” was used. 1) “Mt. Fuji is superb.”
2) “The scenery is terrible. Only the next hotel can be seen.”
3) “The food was bad.”
4) “I want to eat again.”
In 1), “superb view” corresponds to the highly evaluated neighborhood word for “Yabai” in 1), and “invisible” corresponds to the low evaluation neighborhood word for “Yabai” in 2). As for 3) and 4), “bad” and “want to eat” correspond to the low evaluation neighborhood word and the high evaluation neighborhood word for “Yabai”, respectively.
In this case, the “neighborhood” of the high / low evaluation difficult word is, for example, a range within a predetermined number of characters from the high / low evaluation difficult word, a range of one sentence including the high / low evaluation difficult word, or For example, a range of a total of three sentences including one sentence including the high / low evaluation difficult word and one sentence before and after the sentence may be used.
The “neighboring” range may be set as appropriate according to the actual embodiment. At this time, as the range of “neighborhood”, a different range may be set according to the high / low evaluation difficult word. Alternatively, different ranges may be set according to evaluation items.
It is also assumed that different words are set for each evaluation item as the high evaluation neighborhood word and the low evaluation neighborhood word. For example, for a high / low evaluation difficult word “X”, in evaluation item A (eg, “meal”), “Y” (eg, “I want to eat again”) is used as the high evaluation neighborhood word of the high / low evaluation difficult word. ) Is used, but in the evaluation item B (for example, “bath”), when it is assumed that the “Y” is hardly used, the evaluation item B is in the vicinity of “X”. Thus, the process of determining the presence or absence of “Y” is useless.
For this reason, in the storage information of the difficult-to-determine word DB 32 shown in FIG. 7, the high evaluation neighborhood word and the low evaluation neighborhood word may be individually stored for each evaluation item.

When a high / low evaluation difficult word is detected in the input sentence, the upper / lower limit value setting processing unit F1 sets the high / low evaluation difficult word as a target in the vicinity of the high / low evaluation difficult word. The presence / absence of correlated high evaluation neighborhood words and low evaluation neighborhood words is determined. If there is a highly evaluated neighborhood word, an estimation result is obtained that the detected high / low evaluation difficult word is a high evaluation word, and conversely, if there is a low evaluation neighborhood word, the detected high / low The estimation result that the evaluation difficult word is a low evaluation word is obtained.

Although illustration is omitted here, the adjustment value Va is also set for the high / low evaluation difficult word. In the case of this example, the adjustment value Va of the high / low evaluation difficult word is, for example, “10” when the high / low evaluation difficult word is estimated as a high evaluation word, and when it is estimated as a low evaluation word. Is set to “−10”.
Of course, a different adjustment value Va can be set according to the type of high / low evaluation difficult word.

When the high / low evaluation difficult word is detected, the upper / lower limit setting processing unit F1 of this example uses the high / low evaluation difficult word and the adjustment value Va to set the upper limit value and the lower limit value. Set up.

Returning to FIG. 3, when the evaluation input restriction processing unit F2 accepts input of the evaluation value from the user, the evaluation value input permission range is set to a range from the lower limit value to the upper limit value set by the upper and lower limit value setting processing unit F1. In addition to the restriction, information is presented to prompt the user to correct the input sentence in response to an operation input that instructs the input of the evaluation value outside the input permission range.

FIG. 8 shows an example of a web page wp that is presented in response to the operation of the input completion button B1 described above and in which an evaluation value can be input.
As shown in the figure, in the web page wp in this case, the evaluation value input permission range is indicated by graying out stars that cannot be input in the evaluation value input field Rv (in the example of the figure, “2-4”). Indicates the permitted input range). By displaying such an input permission range, the range of evaluation value input by the user is limited.
In this example, message information such as “Please enter the number of ☆” is presented to the position corresponding to the position of the evaluation value input field Rv in order to prompt the user to input the evaluation value.

In addition, in the web page wp in this case, the display for limiting the input range of the evaluation value is performed as described above, and the information presentation for specifying the high evaluation word and the low evaluation word in the input sentence is also performed. .
Specifically, in the input sentence in the evaluation sentence input field Rt, the specified high evaluation word and low evaluation word are highlighted, for example, by surrounding them with a frame.
With such a display, it becomes possible to suggest to the user the part that is the basis for setting the upper limit value and the lower limit value in the input sentence.

The evaluation input restriction processing unit F2 performs different processing on the evaluation value input field Rv shown in FIG. 8 depending on whether the user inputs an evaluation value within the range or inputs an evaluation value outside the range.
FIG. 9 shows an example of the web page wp when the evaluation value is input within the input permission range. As shown in the figure, when the evaluation value input is performed within the input permission range (in the example of the figure, the case where the evaluation value = “3” is input), the evaluation input restriction processing unit F2 is described above. A posting button B2 is presented instead of the input completion button B1. The posting button B2 is a button for instructing the service providing server 30 to post the evaluation text and evaluation value input by the user.

In response to the posting button B2 being operated, the service providing server 30 performs a process of registering input information for the evaluation text input field Rt and the evaluation value input field Rv in the review DB 33 as review information.

FIG. 10 shows an example of stored information in the review DB 33.
The review information is registered by associating the review target ID, the review ID, the contributor ID, and the rejection presence / absence identifier. The review target ID is an identifier that indicates the type of review target, and corresponds to the ID of the accommodation facility in this example. The review ID is an identifier that indicates the type of review, and is assigned by the service providing server 30 for each posted review. The contributor ID is a user ID assigned to the user who has submitted the review. The refusal presence / absence identifier is an identifier indicating whether or not a review is posted after the review text is rejected as will be described later. For example, “1” indicates that the review has been rejected. “0” indicates that the review is not rejected.

In response to the operation of the posting button B2 on the web page wp, the service providing server 30 determines the ID of the facility that is the evaluation target of the web page wp (review posting page) and the user ID of the user who performed the review posting. The ID of the facility (that is, the review target ID) and the user ID (contributor ID), the review ID, the rejection presence / absence identifier, the evaluation value input in the evaluation value input field Rv, and the evaluation text input field Rt The review text input to is stored in the review DB 33 in association with the review text.
In this case, since an evaluation value within the input permission range is input, “0” is stored as the rejection presence / absence identifier.

On the other hand, when there is an operation for instructing the evaluation value input outside the input permission range, the evaluation input restriction processing unit F2 performs an alert display Da1 as shown in FIG. 11 for the user. The alert display Da1 presents information indicating a correction required point in the input sentence for enabling the evaluation value input with the value specified by the input operation on the evaluation value input field Rv.
Specifically, in the alert display Da1, an evaluation value designated by the user (hereinafter, also abbreviated as “desired value” in the sense of an evaluation value desired to be input) is a value larger than the input permission range as the above-mentioned correction required portion. Depending on whether there is a small value or not, a different point is pointed out.
FIG. 11 shows an example corresponding to a case where the desired value by the user is a value (“5”) larger than the input permission range. Since the desired value is larger than the input permission range, in this case, a request is made in the input sentence that contains a low evaluation word (in the example shown in the figure, “cold”, “bad”, “dirty”). This is pointed out as a correction. On the contrary, if the desired value is smaller than the input permission range, a point including a high evaluation word in the input sentence is pointed out as a request correction point.

In this example, the alert display Da1 is displayed by a modal window. As shown as the correction button B3 and the correction rejection button B4 in the figure, there are also operation buttons for allowing the user to select whether or not to correct the input sentence. It is arranged.

Note that the specific method of the alert display Da1 is not limited to the above-described method using a modal window, and other methods such as display with page transition may be employed.

When the correction button B3 is operated in the alert display Da1 as described above, the evaluation input restriction processing unit F2 ends the alert display Da1, presents a web page wp as shown in FIG. To correct. Specifically, for example, message information such as “Please correct” is presented while taking over the input contents in the evaluation text input field Rt.
In addition, in the web page wp shown in FIG. 12, the input completion button B1 is presented instead of the posting button B2.

In this example, when the user corrects the input sentence, if necessary, an assist function for correction, that is, another evaluation value input with the same value as the value (desired value) specified in the evaluation value input operation is permitted. Provides a function of presenting input text to the user.
Specifically, the evaluation input restriction processing unit F2 determines whether or not the character input to the evaluation text input field Rt is stagnant in the presentation state of the web page wp for correction shown in FIG. The user is presented with another input sentence that is permitted to input an evaluation value at the same value as the desired value in accordance with the determination that the user is doing. Whether the character input is stagnant is determined, for example, based on whether the state in which no character input is performed continues for a predetermined time or more.

For example, the “other input text” is presented in a manner as shown in FIG.
Specifically, in this case, the reference display Ds by the modal window is performed while being superimposed on the web page wp shown in FIG. In this reference display Ds, “other input text” is presented.
At this time, “other input text” searches and presents review information satisfying the condition that at least an evaluation value equivalent to the desired value is associated with the stored information in the review DB 33.
Note that as a search condition in this case, a condition of post review by another user (a user other than the user who inputs the desired value) may be added. Since it is possible to refer to other people's ideas, it is possible to realize more useful information presentation when correcting input sentences.
Further, as a search condition, a condition of post review for the same facility (same review target) or a similar facility (similar review target: for example, a facility in the same price range, the same region, etc.) may be added. By presenting the evaluation texts of reviews with the same or similar review targets, it is possible to present to the user information that is more useful for text correction.
When there are a plurality of posted reviews that have been searched for, only a predetermined number of reviews need be displayed for reference.

In the reference display Ds, a display end button B5 is arranged, and the user can end the reference display Ds by operating the display end button B5 to shift to a state in which text can be input to the evaluation text input field Rt. it can.

Although illustration is omitted, when the input completion button B1 is operated after the sentence correction, the service providing server 30 performs the same processing as the above-described upper and lower limit value setting processing unit F1 to perform the upper limit value and the lower limit for the corrected evaluation sentence. Set the value.
Then, when an upper limit value and a lower limit value are set for the modified evaluation text, in this case, it is determined whether or not the desired value instructed by the user is within the input permission range based on the upper limit value and the lower limit value. If it is within the range, a process of registering the desired value and the corrected evaluation text in the review DB 33 according to the operation of the posting button B2 is performed.
On the other hand, if the desired value does not fall within the input permission range, the above-described alert display Da1 is performed again, and if the correction button B3 is operated on the alert display Da1, the web page wp for correcting the sentence is displayed. When the stagnation is determined, the reference display Ds is performed. This series of processing is repeated as long as the desired value does not fall within the input permission range set according to the corrected text and the correction button B3 is continuously operated on the alert display Da1 performed accordingly.

Next, a process when the correction rejection button B4 is operated on the alert display Da1 will be described.
The modification rejection acceptance processing unit F3 illustrated in FIG. 3 accepts a modification rejection operation by the modification rejection button B4. When the correction rejection button B4 is operated, the service providing server 30 first updates the display content of the web page wp as shown in FIG. That is, the web page wp in a state where the user's desired value is displayed in the evaluation value input field Rv and the posting button B2 is arranged instead of the input completion button B1 is presented.

In the state shown in FIG. 14, when the posting button B <b> 2 is operated, the service providing server 30 performs review information registration processing in the review DB 33. However, in this case, since correction of the review text is rejected, “1” is stored as the above-described rejection presence / absence identifier.

When the correction rejection button B4 is operated, the reference learning processing unit F4 shown in FIG. 3 performs a process for correcting the setting reference for the upper limit value and the lower limit value.
The reference learning processing unit F4 analyzes the input sentence for which the rejection operation is received by the correction rejection receiving processing unit F3, and corrects the upper and lower limit value setting processing unit F1 for setting the upper limit value and the lower limit value.

In this example, the correction of the setting criteria for the upper limit value and the lower limit value is performed in response to the number of rejections for an input sentence including the same high evaluation word or low evaluation word being performed a plurality of times.
At this time, in order to be able to count the number of rejections for input sentences including the same high evaluation word or low evaluation word, the reference learning processing unit F4 stores the rejection sentence in the rejection review DB 34 every time the correction of the input sentence is rejected by the user. Information related to reviews for which correction is rejected is stored.

FIG. 15 shows an example of stored information in the rejection review DB 34.
The reference learning processing unit F4, in response to the fact that a new rejection review in which the correction of the input sentence is rejected by the user (the correction rejection button B4 is operated) is registered in the review DB 31, “rejection review ID” in the figure. ”As the review ID of the rejection review, and“ representation word for rejection review ”and“ word type ”as information indicating the evaluation word and high evaluation word / low evaluation word detected from the input text of the rejection review, respectively. Is stored. Further, in association with these pieces of information, as the “desired value deviation type” in the figure, the user's desired value for the rejection review is larger (larger deviation) or smaller (smaller deviation). ) Is stored.

Here, in this example, since it is estimated as the high / low evaluation word that the high / low evaluation determination difficult word is used, the high evaluation word and the low evaluation word stored in the rejection review DB 34 as described above. (In other words, high evaluation words and low evaluation words that are subject to correction of the upper and lower limit setting criteria) include high / low evaluation determination difficult words.
That is, when there is a high / low evaluation determination difficult word estimated as “high evaluation word” or “low evaluation word” during the rejection review, the reference learning processing unit F4 regards the “high evaluation word” and “low evaluation word”. Is also stored in the rejection review DB 34.

As described above, information associated with “rejection review ID”, “evaluation word during rejection review”, “word type”, and “desired value deviation type” that is updated each time a new rejection review occurs is referred to as “rejection review management”. It is written as “Information I1”.

In addition, when a new rejection review is registered, the reference learning processing unit F4 performs a process for updating the word rejection count management information I2 as shown in FIG. The word rejection count management information I2 is information associating an evaluation word being rejected with a count value as shown in the figure. In the case of this example, the word rejection count management information I2 is also stored in the rejection review DB 34.
The reference learning processing unit F4, when a new refusal review is registered and the refusal review management information I1 is updated, an evaluation word that matches the evaluation word in the refusal review in the word refusal count management information I2 Is present, the count value corresponding to the evaluation word is incremented by one. When there is no evaluation word stored in the word rejection count management information I2 in the rejection review, the reference learning processing unit F4 newly adds the corresponding evaluation word in the rejection review to the word rejection count management information I2, “1” is stored as the count value corresponding to the added evaluation word.
In this example, for a certain evaluation word “W” stored in the word rejection number management information I2, even if there are a plurality of the evaluation words “W” in a new rejection review, the evaluation word “W” The count-up amount corresponding to “” is “1”.

The reference learning processing unit F4 uses the rejection review management information I1 and the word rejection count management information I2 as described above to correct the setting criteria for the upper limit value and the lower limit value.
Specifically, when updating the word rejection count management information I2 based on a new rejection review, it is determined whether there is an evaluation word having a count value equal to or greater than a predetermined value (at least a natural number of 2 or more). If there is a corresponding evaluation word, the adjustment value Va for the evaluation word is corrected.
Here, the determination of “whether there is an evaluation word having a count value equal to or greater than a predetermined value” as described above is based on whether the number of rejections for an input sentence including the same high evaluation word or low evaluation word is a predetermined number of times. This is equivalent to determining whether or not the above has been reached.

At this time, the adjustment value Va is corrected based on the small side deviation (desired value <input permission range), the large side as the relationship between the desired value and the input permission range when the correction of the review including the corresponding evaluation word is rejected. This is based on which of the deviations (desired value> input permission range) is dominant (that is, which of the number of rejection reviews due to minor deviations and the number of rejection reviews due to major deviations is greater). Whether the minor deviation or the major deviation is dominant can be determined based on the desired value deviation type information in the rejection review management information I1 shown in FIG.

Hereinafter, the case where the corresponding evaluation word is a high evaluation word and the case where it is a low evaluation word will be specifically described.
When the corresponding evaluation word is a high evaluation word, if rejection due to a small deviation is dominant, the adjustment value Va for the high evaluation word is estimated to be larger than expected, and therefore the adjustment value Va is made smaller. To correct. On the other hand, if the rejection due to the large deviation is dominant, the adjustment value Va for the high evaluation word is smaller than expected, so that the adjustment value Va is corrected to be larger.
When the threshold value for the count value (the predetermined value described above) is an even number, the number of rejections due to small deviation can be equal to the number of rejections due to large deviation. Therefore, it is desirable to set the predetermined value to an odd number.

Further, when the corresponding evaluation word is a low evaluation word, if rejection due to a small deviation is dominant, the adjustment value Va for the low evaluation word is larger than expected (as an absolute value). Therefore, the adjustment value Va is corrected so as to be smaller. On the other hand, if the rejection due to the large deviation is dominant, the adjustment value Va for the low evaluation word is smaller than expected (it was large as an absolute value), so the adjustment value Va is corrected to be larger. .

Regarding the correction of the adjustment value Va as described above, the correction amount of the adjustment value Va may be fixed, or the degree of dominance of the minor deviation and the major deviation (the number of rejection reviews due to the minor deviation and the number of rejection reviews due to the major deviation) It is also possible to make it variable according to the difference between the evaluation word and the corresponding evaluation word.
If the correction amount of the adjustment value Va is fixed, the burden of the correction process can be reduced, and if it is variable, the adjustment value Va can be finely corrected according to various conditions, and the setting accuracy of the upper limit value and the lower limit value can be further increased. Improvements can be made.

In the above description, regarding the addition of the count value in the word rejection count management information I2 in response to the occurrence of a new rejection review, even if there are a plurality of corresponding evaluation words existing in the rejection review, the count value is “ Although an example in which only “1” is added has been given, the count value may be added by the same number as the number of corresponding evaluation words in the rejection review. In that case, even if only one user rejects the correction for the corresponding evaluation word, the adjustment value Va for the evaluation word can be corrected.

[1-4. Processing procedure]

Next, a specific processing procedure to be executed for realizing the function of the service providing server 30 as the first embodiment described above will be described with reference to the flowcharts of FIGS.
The processes shown in FIGS. 17 to 19 are executed by the CPU 101 of the service providing server 30 in accordance with a program stored in a required storage device such as the ROM 102, for example.

17 and 18 show processing from the input of review information to the registration of review information.
In FIG. 17, the service providing server 30 (CPU 101) waits for a page request, that is, a request for a web page wp as a review posting page in step S101. A request for a review posting page for a certain facility is made by the user clicking a button provided on a predetermined web page in the service providing site as described above.
Based on the page request, the service providing server 30 can specify the facility (product) to be evaluated and the user ID of the review posting user.

If there is a page request in step S101, the service providing server 30 transmits the web page wp requested in step S102 to the requesting user terminal 4. As a result, the web page wp in the initial state as shown in FIG. 4 is displayed on the user terminal 4, and the user can input the evaluation text in the evaluation text input field Rt in the web page wp.

In subsequent step S103, the service providing server 30 waits for an input completion instruction for the evaluation text by the input completion button B1, and when the input completion instruction is received, an evaluation word detection process is performed in step S104. The evaluation word detection process includes the above-described high / low evaluation difficult word detection process. If a high / low evaluation difficult word is detected, an estimation process as to whether the word is a high evaluation word or a low evaluation word is also performed.
In addition, the detection method of the high evaluation word, the low evaluation word, and the high / low evaluation determination difficult word from the input sentences based on the stored information of the evaluation word DB 31 and the difficult determination word DB 32, and the high evaluation in the difficult determination word DB 32 Since the estimation method of the high evaluation word / low evaluation word for the high / low evaluation determination difficult word based on the neighborhood word and the low evaluation neighborhood word has already been described, redundant description is avoided.

In response to the execution of the evaluation word detection process in step S104, the service providing server 30 performs an upper limit value and lower limit value setting process in step S105.
In addition, since the setting method of the upper limit value and the lower limit value using the adjustment value Va based on the detected evaluation word has already been described, redundant description is avoided.

In subsequent step S106, the service providing server 30 performs processing for activating the evaluation value input field Rv and presenting the basis information. That is, as shown in FIG. 8 above, the evaluation value input field Rv in the web page wp is activated (enabling the evaluation input) in the state where the input permission range by the upper limit value and the lower limit value set is shown, In the evaluation text input field Rt, processing for highlighting the evaluation words (including difficult-to-determine words) detected by the above detection processing is performed.

In further subsequent step S107, the service providing server 30 waits for an evaluation value input operation for the evaluation input field Rv. If there is an evaluation value input operation, it is determined in step S108 whether the value is within the input permission range. . That is, it is determined whether or not the evaluation value instructed to be input by the evaluation value input operation is a value within the input permission range.

If it is determined in step S108 that the value is within the input permission range, the service providing server 30 proceeds to step S109 and performs processing for switching the input completion button B1 on the web page wp to the posting button B2 (see FIG. 9). ). At this time, the evaluation value instructed to be input is displayed in the evaluation value input field Rv (evaluation value = 3 in the example of FIG. 9).

In step S126, the service providing server 30 waits for the posting operation, that is, the operation of the posting button B2, and when there is a posting operation, the review information registration processing is performed in step S127, and a series of processes shown in FIGS. Finish.
Note that the registration process of step S127, that is, the process of registering the review information in the review DB 33 when the evaluation value is input within the input permission range has already been described, and therefore a duplicate description is avoided.

If it is determined in step S108 that it is not within the input permission range, the service providing server 30 advances the process to step S110 shown in FIG.
In step S110 illustrated in FIG. 18, the service providing server 30 performs a process for identifying a portion requiring correction. As described above, the point requiring correction here means a point requiring correction in the input sentence so that the user can input the evaluation value with the desired value, and the desired value is within the input permission range. The different parts in the input sentence are specified depending on whether the deviation is a minor deviation or a major deviation. In addition, since the specific specification method of the correction required part has been demonstrated, duplication description is avoided.

In step S111 following step S110, the service providing server 30 performs an alert display Da1 indicating a correction required location (see FIG. 11), and in further subsequent step S112, the buttons (correction button B3, correction rejection button B4) in the alert display Da1. ) Wait for operation.

If there is a button operation in step S112, the service providing server 30 determines whether the correction button B3 is operated in step S113.
When the correction button B3 is operated, the service providing server 30 proceeds to step S114 to perform processing for ending the alert display Da1 (see FIG. 12), and the evaluation text input field Rt is activated in step S115. Wait until. Activation as used herein means that the user has performed an operation of transitioning the evaluation text input field Rt to a character input enabled state. In this example, for example, the evaluation text input field Rt is clicked or tapped. Means that the operation has been performed.

In response to the evaluation text input field Rt being activated, the service providing server 30 executes stagnation detection processing in step S116. That is, a process for detecting that the text input for correction is in a stagnation state is performed. In the stagnation detection process in this example, the time length during which the character input undetected state continues after the evaluation text input field Rt is activated is counted, and when the time length reaches a predetermined time length, the stagnation flag is set. It is set to “1” (the initial value is “0”), and the process proceeds to step S117. On the other hand, if a character input is detected before the time length that has been measured reaches the predetermined time length, the stagnation flag is kept “0” and the process proceeds to step S117.

In step S117, the service providing server 30 determines whether or not it is stagnant, that is, whether or not the above stagnant flag is “1”.

Here, as the stagnation determination process, it is determined only whether or not the first character input in the evaluation text input field Rt is stagnation. However, for example, the stagnation determination process is performed for each character input or item. For example, the first character of each character may be targeted, and the first character in the evaluation text input field Rt may be targeted.

If it is determined in step S117 that it is stagnant, the service providing server 30 performs processing for reference display of other reviews in step S118. That is, as described above as the reference display Ds in FIG. 13, the process for causing the user to be presented with information including the input text of another review in which the evaluation value input with the same value as the desired value is permitted. I do. Note that a specific method for performing the reference display Ds has already been described, and thus a duplicate description is avoided.

In step S119, the service providing server 30 waits for a closing operation on the reference display Ds, that is, an operation of the display end button B5. If there is a closing operation, the service providing server 30 ends the reference display Ds in step S120, and the process proceeds to step S121. Proceed.

If it is determined in step S117 that there is no stagnation, the service providing server 30 passes the above steps S118 to S120 and proceeds to step S121.

In step S121, the service providing server 30 stands by until the operation of the input completion button B1 is performed. In response to the operation, the service providing server 30 performs an evaluation word redetection process in step S122. That is, the evaluation word detection process is performed by the same method as in the previous step S104 for the input sentence in the evaluation sentence input field Rv determined by the operation of the input completion button B1 in step S121.

Further, in the subsequent step S123, the service providing server 30 sets the upper limit value and the lower limit value by the same method as in the previous step S106 based on the evaluation word detected in step S122, as the upper limit value and lower limit value resetting process. Perform the process.

In response to the execution of the resetting process in step S123, the service providing server 30 determines in step S124 whether the desired value is within the input permission range, that is, whether the desired value is within the range from the reset lower limit value to the upper limit value. Determine whether or not.
If the desired value is not within the input permission range, the service providing server 30 returns to the previous step S110. In other words, as long as the desired value does not fall within the input permission range set according to the corrected text and the correction button B3 is continuously operated in the alert display Da1 performed accordingly, the input text can be corrected. The process for resetting and the upper limit value and the lower limit value based on the corrected text are repeated.

On the other hand, if the desired value is within the input permission range, the service providing server 30 performs processing for switching the input completion button B1 on the web page wp to the posting button B2 in step S125, and then step S126 shown in FIG. Proceed with the process.
As a result, when the input sentence is corrected and the desired value falls within the input permission range, the review information registration process of step S127 is performed according to the operation of the posting button B2.

Subsequently, when it is determined in the previous step S113 that the button on which the operation has been performed is not the correction button B3, the service providing server 30 proceeds to step S128, and changes the input completion button B1 on the web page wp to the posting button B2. After performing the process for switching, it waits until posting operation is performed by step S129.
If a posting operation is performed, the service providing server 30 sets the correction rejection flag to ON (“1”) in step S130, and performs review information registration processing for the review DB 33 in step S131. In the review information registration process in this case, in response to the correction rejection flag being ON, “1” indicating that the review has been rejected for correction is stored as a rejection presence / absence identifier (see FIG. 10).

Further, in the subsequent step S132, the service providing server 30 performs information registration processing in the rejection review DB 34. That is, as described above as the rejection review management information I1, the information of “rejection review ID”, “evaluation word during rejection review”, “word type”, and “desired value deviation type” is stored in the rejection review DB 34 in association with each other. Process.

In response to the execution of the information registration process in step S132, the service providing server 30 finishes the series of processes shown in FIGS.

In the above example, the case where the user takes the evaluation value input outside the input permission range is exemplified as the case where only the correction of the input sentence is taken, but of course the evaluation value is not limited to the correction of the input sentence. It is also possible to allow the user to select reinsertion.

FIG. 19 shows a process for learning the setting criteria for the upper limit value and the lower limit value.
In FIG. 19, the service providing server 30 stands by in step S201 until a new refusal review is registered (registration with respect to the review DB 31). If a refusal review is registered, an evaluation word under review is added in step S202. get.

In response to the acquisition of a new evaluation word during rejection review in step S202, the service providing server 30 updates the word rejection count management information I2 in step S203. When the word rejection count management information I2 includes an evaluation word that coincides with an evaluation word that is in the new rejection review, the word rejection count management information I2 counts corresponding to the evaluation word in the word rejection count management information I2. When the evaluation word stored in the word rejection count management information I2 does not exist in the rejection review, the corresponding evaluation word in the rejection review is newly added to the word rejection count management information I2. , “1” is stored as a count value corresponding to the added evaluation word.

In subsequent step S204, the service providing server 30 determines whether there is an evaluation word having a count value ≧ predetermined value.
If there is an evaluation word with a count value ≧ predetermined value, the service providing server 30 proceeds to step S205, performs a process of correcting the adjustment value Va of the corresponding word, and ends the process shown in FIG. Note that the correction process for the adjustment value Va in this example is performed based on whether the rejection due to the minor deviation or the rejection due to the major deviation is dominant as the modification rejection for the review including the corresponding evaluation word. Since a specific method has already been described, a duplicate description is avoided.
On the other hand, if there is no evaluation word satisfying the count value ≧ predetermined value, the service providing server 30 passes the correction process of step S205 and ends the process shown in FIG.

[1-5. Modifications related to adjustment value correction trigger]

In the above, the case where the adjustment of the adjustment value Va, that is, the correction of the setting criterion for the upper limit value and the lower limit value, is performed according to only the number of rejections for input sentences including the same high evaluation word or low evaluation word.
Alternatively, the setting criteria may be modified in response to rejection of input sentences including the same high evaluation word or low evaluation word by a plurality of users having different attributes.

In this case, in the information registration process for the new rejection review with respect to the rejection review DB 34, as shown in FIG. 20, the information of “rejection review ID”, “evaluation word during rejection review”, “word type”, and “desired value deviation type” is added. In addition, “poster attribute” information is stored in association with each other.
Here, as an attribute of a contributor (user), for example, age group, gender distinction, residence area distinction, and the like can be cited. Alternatively, when analyzing the purchase behavior of products for each user, use the genre of products that are often purchased, the price range of products that are often purchased, the frequency of product purchase (site usage frequency), etc. as attribute information. Is also possible.
In this case, the service providing server 30 acquires attribute information corresponding to the poster of the rejection review and stores it in the rejection review DB 34 when registering information about the new rejection review.

FIG. 21 is a flowchart showing a procedure of processing to be executed in the modification.
Note that the service providing server 30 in this case also executes the processing shown in FIGS. In this case, the service providing server 30 executes the process shown in FIG. 21 instead of the process shown in FIG.
In the following description, the same processes as those already described are denoted by the same reference numerals and the description thereof is omitted.

In FIG. 21, when it is determined that there is an evaluation word having a count value ≧ predetermined value in step S204, the service providing server 30 in this case is rejected by a user having a different attribute in the corresponding word in step S301. It is determined whether or not there is a word.
That is, for an evaluation word with a count value ≧ predetermined value, it is determined whether or not the evaluation word is rejected by a user with a different attribute based on the information stored in the rejection review DB 34 shown in FIG.

If there is a corresponding evaluation word, the service providing server 30 performs the correction process in step S205 and ends the process shown in FIG.
On the other hand, if there is no corresponding evaluation word, the service providing server 30 passes the correction process of step S205 and ends the process shown in FIG.

As described above, if the condition that the rejection of the input sentence containing the same evaluation word was made by a plurality of users with different attributes is set as the correction trigger condition of the setting standard, the setting standard is erroneously set. Corresponding to the case where it is estimated that there is a higher possibility that the user has been set, the setting standard can be corrected.

Note that the modification described above with reference to FIGS. 20 and 21 can also be applied to a second embodiment described later.

[1-6. Summary of First Embodiment]

As described above, the information processing apparatus (service providing server 30) as the first embodiment analyzes an input sentence by a user, and calculates an evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence. Upper / lower limit value setting unit (upper / lower limit value setting processing unit F1) for setting upper limit value and lower limit value, and upper / lower limit value setting unit set the input permission range of evaluation value when receiving input of evaluation value from user An evaluation input restriction unit (evaluation input restriction) that restricts the range from the lower limit value to the upper limit value and presents information that prompts the user to correct the input sentence according to the operation input that instructs the input of the evaluation value outside the permitted input range And a processing unit F2).
Further, the correction rejection receiving unit (correction rejection receiving processing unit F3) that receives the correction rejection operation, and the input sentence in which the rejection operation is received by the correction rejection receiving unit are analyzed, and the upper limit value and the lower limit value by the upper and lower limit value setting unit And a reference learning unit (reference learning processing unit F4) for correcting the setting reference.

According to the said structure, the range of the evaluation value in which the input by a user is permitted is restrict | limited according to the content of an input sentence. That is, it is possible to prevent mistakes in evaluation value input while leaving room for the user to set evaluation values.
Therefore, it is possible to prevent a discrepancy between the content of the evaluation text and the evaluation value under the assumption that the user himself inputs the evaluation value, and to improve the reliability of the evaluation information.
Further, the learning function based on the analysis of the input sentence whose correction is rejected improves the setting accuracy of the upper limit value and the lower limit value, which also contributes to the improvement of the reliability of the evaluation information.

In the information processing apparatus according to the present embodiment, the reference learning unit corrects the setting criterion according to the number of rejections for input sentences including the same high evaluation word or low evaluation word.

Thereby, it is possible to prevent the setting criteria for the upper limit value and the lower limit value from fluctuating due to noise when only one user rejects the correction.
Therefore, it is possible to improve the setting accuracy of the upper limit value and the lower limit value, and contribute to further improvement of the reliability of the evaluation information.

Furthermore, in the information processing apparatus according to the present embodiment, the reference learning unit corrects the setting criterion, and rejection of input sentences including the same high evaluation word or low evaluation word is performed by a plurality of users having different attributes. Depending on the situation.

Thereby, it is possible to correct the setting standard corresponding to a case where it is estimated that there is a higher possibility that the setting standard is incorrect.
Therefore, it is possible to further improve the setting accuracy of the upper limit value and the lower limit value and contribute to further improvement of the reliability of the evaluation information.

Furthermore, in the information processing apparatus according to the present embodiment, the upper and lower limit value setting unit has a meaning of either a high evaluation word or a low evaluation word from a nearby context for a high / low evaluation determination difficult word detected from an input sentence. It is used to set the upper and lower limit values as a word meaning the estimated one.

This makes it possible to increase the number of samples of high evaluation words and low evaluation words required for setting the upper limit value and the lower limit value.
As the number of samples increases, it is possible to improve the setting accuracy of the upper limit value and the lower limit value, which can contribute to further improving the reliability of the evaluation information.

Further, in the information processing apparatus of the present embodiment, the evaluation input restriction unit presents information for specifying the high evaluation word and the low evaluation word in the input sentence to the user.

As a result, it is possible to suggest to the user the part that is the basis for setting the upper limit value and the lower limit value in the input sentence.
Therefore, sentence correction for inputting a desired evaluation value can be facilitated.

Furthermore, in the information processing apparatus of the present embodiment, the evaluation input restriction unit is designated by the input operation when an input operation for designating a value outside the input permission range is performed as the evaluation value input operation. The user is presented with information that points out the points that need to be corrected in the input text to enable the evaluation value input by value.

Thereby, since the input of the desired evaluation value is permitted, it is possible to suggest to the user which part in the input sentence should be corrected.
Therefore, it is possible to further facilitate sentence correction for inputting a desired evaluation value.

Furthermore, in the information processing apparatus according to the present embodiment, the evaluation input restriction unit is designated by the input operation when an input operation for designating a value outside the input permission range is performed as the evaluation value input operation. The user is presented with other input sentences that are permitted to be input with the evaluation value at the same value.

Thus, it is possible to suggest to the user what kind of sentence correction should be performed in order to allow the input of the desired evaluation value.
Therefore, it is possible to further facilitate sentence correction for inputting a desired evaluation value.

<2. Second Embodiment>
[2-1. Function and processing procedure]

Next, a second embodiment will be described.
In the second embodiment, the configuration of the network system 1 including the service providing server 30 is the same as that in the first embodiment, and a description thereof will be omitted. In the following description, differences from the first embodiment will be mainly described.

FIG. 22 shows an example of a web page wp ′ presented as a review posting page in the second embodiment.
In the second embodiment, the input form of evaluation information is different from that of the first embodiment. Specifically, an evaluation value can be input for each evaluation item, and a comprehensive evaluation value (hereinafter referred to as “total evaluation value”) for an evaluation target can be input.
As shown in the figure, in the web page wp ′, as the evaluation column Rk for each evaluation item, in this case, five evaluation columns Rk1 to Rk5 corresponding to the respective items of “bath”, “meal”, “room”, “location”, and “service” And a comprehensive evaluation value input field Rsv for inputting a comprehensive evaluation value.
In each evaluation field Rk (Rk1 to Rk5), an item evaluation value input field Rkv (indicated as Rkv1 to Rkv5 for each item) for inputting an item evaluation value and an evaluation text for the item are input. Item evaluation text input field Rkt (represented as Rkt1 to Rkt5 for each item).

The display content of the web page wp ′ shown in FIG. 22 is the display content in the initial state before input. In the web page wp ′ in the initial state, the star mark is not grayed out in the item evaluation value input fields Rkv1 to Rkv5, and is activated in a state where an arbitrary evaluation value can be input.
In the second embodiment, as a procedure up to review posting, it is premised that first, a sentence and an evaluation value are input for each evaluation item, and then an overall evaluation value is input. For this reason, in the initial state web page wp ′, all the star marks in the comprehensive evaluation value input field Rsv are grayed out, and the evaluation value input is disabled.
Further, in order to correspond to the above procedure, an item evaluation completion button B6 for instructing that the item evaluation is completed is presented on the web page wp ′ in the initial state.

FIG. 23 illustrates a state in which text input and evaluation value input for each evaluation item are performed on the web page wp ′ in the initial state.
In this way, the user inputs a sentence and an evaluation value for each evaluation item, operates the item evaluation completion button B6, and inputs a comprehensive evaluation value.

FIG. 24 shows a display content example of the web page wp ′ presented in response to the operation of the item evaluation completion button B6.
As can be seen by referring to the comprehensive evaluation value input field Rsv in the figure, in the second embodiment, an evaluation value input permission range is set for the comprehensive evaluation value. Specifically, the upper and lower limit value setting processing unit F1 in this case sets an upper limit value and a lower limit value for the overall evaluation value according to the number of high evaluation words and low evaluation words in the entire input sentence over a plurality of evaluation items. .
In this case as well, the method for setting the upper limit value and the lower limit value is the same as in the case of the first embodiment, and a duplicate description is avoided.

Also in this case, in response to the input permission range being indicated, the information for specifying the high evaluation word and the low evaluation word is indicated in the input sentence as in the first embodiment (upper and lower limits and Presenting the grounds for setting the lower limit).

FIG. 25 exemplifies the state of the web page wp ′ when an operation for instructing an evaluation value input (for example, “4”) within the input permission range is performed for the comprehensive evaluation value.
In the web page wp ′ in this case, a posting button B2 is presented instead of the previous item evaluation completion button B6. Also in this case, the service providing server 30 performs a process of registering (storing) the input information in the review DB 33 in response to the operation of the posting button B2.
Although illustration is omitted, for the review DB 33 in this case, the evaluation value and the comprehensive evaluation value for each evaluation item are stored as the “evaluation value” shown in FIG. 10, and the “review text” is for each evaluation item. Stores input text.

In addition, although description by illustration is abbreviate | omitted, according to the operation which instruct | indicates the evaluation value input outside an input permission range being performed about the comprehensive evaluation value, also in this case similarly to 1st Embodiment To correct the input sentence. At this time, as in the first embodiment, an alert display Da1 (see FIG. 11) showing a portion requiring correction, or a reference display Ds (FIG. 13) according to the stagnation determination result of the sentence input when the input sentence is corrected. See also).

Here, in the case where the evaluation for each item and the comprehensive evaluation are possible as in the second embodiment, it is estimated that the evaluation value and the text content coincide as the comprehensive evaluation. In each item, there may be a case where there is a discrepancy between the evaluation value and the text content.
FIG. 26 shows an example of wrinkles in item units.
In each item indicated by “O” in the figure, there is a discrepancy between the evaluation value and the text content. Specifically, in the “room” item, a relatively high evaluation value is input as an evaluation value even though the content of the text is low evaluation content. The “location” item is the opposite.
Such wrinkles in item units may be overlooked when viewed from the relationship between the entire input sentence and the overall evaluation value. In particular, as in the example of FIG. 26, in the case where the same number of 齟齬 items of “text content = low evaluation, evaluation value = high evaluation” and 齟齬 items of “text content = high evaluation, evaluation value = low evaluation” occur. It is difficult to detect this in the relationship between the entire input sentence and the overall evaluation value.

Therefore, in the second embodiment, for each evaluation item, the presence / absence of a defect between the input sentence and the input evaluation value is determined, and the evaluation item determined to have a defect is notified to the user. Although not shown, the service providing server 30 according to the second embodiment is provided with a bag item notification processing unit F5 that realizes such a function.

Specifically, the bag item notification processing unit F5 determines the input text and evaluation value for each evaluation item in response to completion of text input and evaluation value input for each evaluation item (in this example, operation of the item evaluation completion button B6). Determine the presence or absence of wrinkles. The determination of the presence or absence of wrinkles is performed based on the relationship between the number of high evaluation words and low evaluation words in the input sentence (including high / low evaluation determination difficult words after estimation) in the input sentence and the input evaluation value. For example, the value of “high evaluation word count” − “low evaluation word count” is calculated, and when the value is positive, if the input evaluation value is 2 or less, there is a defect, whereas conversely it is negative If the input evaluation value is 4 or more, it is determined that there is a defect.

FIG. 27 shows an example of a wrinkle alert display Da3 that is performed in response to the presence of an item determined to have wrinkles.
In this example, the heel alert display Da3 is performed by a modal window in the same manner as the previous alert display Da1. The alert display Da3 includes information in the evaluation column Rk (an input content is also presented) for an item determined to have wrinkles, and a correction button for instructing whether or not to correct the wrinkle item. B7, uncorrected button B8 is presented.

When the correction button B7 is operated, the 齟齬 alert display Da3 is terminated, that is, correction of the input sentence and the evaluation value is accepted on the web page wp ′ in the state shown in FIG.
In this case, when the item evaluation completion button B6 is operated again, the above-described 齟齬 item notification processing unit F5 performs the 齟齬 determination, and if the 齟齬 item exists, the 齟齬 alert display Da3 is displayed again. . In this example, the same process is repeated until the wrinkle is eliminated.

On the other hand, when the wrinkle is resolved, an upper limit value and a lower limit value are set for the comprehensive evaluation value, and the input permission range based on the upper limit value and the lower limit value is set in the comprehensive evaluation value input field Rsv as in FIG. Presented web page wp ′ is presented.

Here, by performing the setting process of the upper limit value and the lower limit value for the comprehensive evaluation value in accordance with the elimination of wrinkles as described above, useless setting processing while the wrinkles are not eliminated is avoided, and the processing burden is reduced. Figured.

FIG. 28 is a flowchart showing a procedure of processing to be executed in order to realize the function as the second embodiment described above.
In this case, the service providing server 30 executes the process shown in FIG. 28 instead of the process shown in FIG.
In this case, the process of step S103 is a process of waiting for the operation of the item evaluation completion button B6 on the web page wp ′, and the evaluation word detection process of step S104 is performed on the entire input sentence over a plurality of evaluation items. .

In this case, the service providing server 30 executes the wrinkle detection process for each item in step S401 in response to the execution of the evaluation word detection process in step S104. Specifically, for each evaluation column Rk in the web page wp ′, a value based on the above “high evaluation word count” − “low evaluation word count” is calculated.

In subsequent step S402, the service providing server 30 determines whether there is a defect item. That is, in this example, the value obtained by calculating “high evaluation word number” − “low evaluation word number” calculated in step S401 is positive while the input evaluation value is 2 or less, or the value is negative. In contrast, it is determined whether there is an evaluation item that satisfies any of the conditions that the input evaluation value is 4 or more.

If there is a bag item, the service providing server 30 proceeds to step S403 to display the bag alert display Da3, and further performs an operation standby process in step S404. That is, the operation waits for either the correction button B7 or the non-correction button B8 on the alert display Da3.
If there is an operation, the service providing server 30 determines whether or not the correction button B7 is operated in step S405. If the operation is the correction button B7, the service providing server 30 proceeds to step S407, ends the alert display Da3, and returns to step S103. . By executing the display end process in step S407, a web page wp ′ based on the display contents at the time when the item evaluation completion button B6 is operated is presented (see FIG. 26), and the user enters the web page wp ′ in this state. On the other hand, the input text in the item evaluation text input field Rkt and the input evaluation value in the item evaluation value input field Rkv can be corrected.
By returning to step S103 as described above, the wrinkle determination for each item is performed again according to the operation of the item evaluation completion button B6 again. If there is a wrinkle item, the wage alert display Da3 is displayed again. Done.

If it is determined in step S405 that the button is not the correction button B7, the service providing server 30 proceeds to step S406, ends the alert display Da3, and then proceeds to step S408.

On the other hand, if it is determined in step S402 that there is no defect item, the service providing server 30 passes the processing of steps S403 to S406 and proceeds to step S408.

In step S408, the service providing server 30 sets the upper limit value and the lower limit value of the comprehensive evaluation value, and in the subsequent step S409, executes processing for activating the comprehensive evaluation value input field Rsv and presenting the basis information. In other words, activation (enabling the evaluation input) is performed in a state where the input permission range based on the upper limit value and the lower limit value in which the comprehensive evaluation value input field Rsv in the web page wp ′ is set is shown, and the detection process (S104) A process for highlighting the evaluation word (including the difficult-to-determine word) detected in (1) is performed (see FIG. 24).

In step S410 following step S409, the service providing server 30 waits for a comprehensive evaluation value input operation. If there is an input operation, the service providing server 30 executes the processing in step S108 and subsequent steps.
Since the processing after step S108 is the same as in the case of the first embodiment, redundant description is avoided.
Also in the second embodiment, the same processing as that shown in FIG. 19 is performed to learn the upper limit and lower limit setting criteria (the adjustment value Va in this example).

[2-2. Summary of Second Embodiment]

As described above, the information processing apparatus according to the second embodiment accepts input text and evaluation value input from the user for each item in the plurality of evaluation items, and separately from the evaluation value for each item. Accepts input of comprehensive evaluation values.
Then, the upper and lower limit value setting unit sets an upper limit value and a lower limit value for the overall evaluation value according to the number of high evaluation words and low evaluation words in the entire input sentence over a plurality of items, and the evaluation input restriction unit For the evaluation value, the limitation of the input permission range and the information presentation that prompts the correction of the input sentence are performed.
Then, for each item, a 齟齬 item notification unit (齟齬 item notification processing unit F <b> 5) that determines whether or not there is a crease between the input sentence and the input evaluation value and notifies the user of the item that is determined to have crease. It has more.

As described above, when the evaluation for each item and the comprehensive evaluation are possible, it is estimated that the evaluation value and the text content match as the comprehensive evaluation. There may also be cases where values and text content are flawed.
According to the above configuration, it is possible to notify the user of the occurrence of wrinkles and their locations on an item basis corresponding to such a case.
Therefore, it is possible to prevent mistakes in evaluation value input and to improve the reliability of evaluation information.

<3. Program and Storage Medium>

As described above, the service providing server 30 as the embodiment of the information processing apparatus according to the present invention has been described. However, the program according to the present embodiment causes the information providing apparatus (CPU or the like) to execute the processing of the service providing server 30. It is a program.

The program according to the embodiment analyzes an input sentence by a user, and sets an upper limit value and a lower limit value of an upper limit value and a lower limit value of an evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence, and When receiving the input of the evaluation value from the user, the input permission range of the evaluation value is limited to a range from the lower limit value to the upper limit value set by the upper and lower limit value setting function, and is outside the input permission range. An evaluation input restriction function for presenting information for urging correction of the input sentence in response to an operation input for instructing input of the evaluation value, a correction rejection reception function for receiving the correction rejection operation, and the correction rejection reception An information processing apparatus that analyzes an input sentence in which the rejection operation is accepted by a function and corrects the upper limit value and the lower limit value setting criterion by the upper and lower limit value setting function To be executed.
That is, this program corresponds to a program that causes the information processing apparatus to execute the processes described with reference to FIGS. 17 to 19, 21, and 28.

With such a program, the information processing apparatus as the service providing server 30 described above can be realized.
Such a program can be stored in advance in an HDD as a storage medium built in a device such as a computer device, a ROM in a microcomputer having a CPU, or the like. Alternatively, it can be stored (stored) temporarily or permanently in a removable storage medium such as a semiconductor memory, memory card, optical disk, magneto-optical disk, or magnetic disk. Such a removable storage medium can be provided as so-called package software.
Further, such a program can be installed from a removable storage medium to a personal computer or the like, or can be downloaded from a download site via a network such as a LAN or the Internet.

<4. Modification>

The present invention is not limited to the specific examples described above, and various modifications can be considered.
For example, in the above description, when the evaluation value input for each evaluation item is enabled, the case where the input restriction of the evaluation value according to the number of high / low evaluation words is performed for the overall evaluation value is exemplified. Can also be performed on the evaluation value for each evaluation item.

In the above, the case where the evaluation target is an accommodation facility has been illustrated, but the present invention can be applied to any evaluation target. For example, tangible and intangible objects that are not premised on sales, such as not-for-sale items, may be set as evaluation targets, not limited to products as sales targets.

Moreover, as for the input form of the evaluation value, other forms such as inputting the numerical value itself can be adopted in addition to the illustrated number of stars.

1 network system, 2 network, 3 service providing system, 30 service providing server, F1 upper / lower limit setting processing unit, F2 evaluation input restriction processing unit, F3 modification rejection acceptance processing unit, F4 reference learning processing unit, F5 input item notification processing Part, wp, wp 'web page, Rv evaluation value input field, Rt evaluation text input field, Rsv comprehensive evaluation value input field, Rk1-5 evaluation field, Rkv1-5 item evaluation value input field, Rkt1-5 item evaluation text input Column, Da1, Da2 alert display, Ds reference display, Da3 齟齬 alert display

Claims

An upper and lower limit setting unit that analyzes an input sentence by a user and sets an upper limit value and a lower limit value of an evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence;
When receiving the input of the evaluation value from the user, the input permission range of the evaluation value is limited to a range from the lower limit value to the upper limit value set by the upper and lower limit value setting unit, and is outside the input permission range. An evaluation input restriction unit for presenting information for prompting correction of the input sentence in response to an operation input instructing input of the evaluation value at
A correction refusal acceptance unit that accepts the refusal operation of the correction;
An information processing apparatus comprising: a reference learning unit that analyzes an input sentence in which the rejection operation is received by the correction rejection receiving unit, and corrects the upper limit value and the lower limit value setting criterion by the upper and lower limit value setting unit.
The reference learning unit
The information processing apparatus according to claim 1, wherein the setting criterion is corrected according to a rejection count for input sentences including the same high evaluation word or low evaluation word.
The reference learning unit
The information processing apparatus according to claim 1, wherein the setting criterion is corrected in response to rejection of input sentences including the same high evaluation word or low evaluation word by a plurality of users having different attributes.
The upper and lower limit value setting unit
For the high / low evaluation difficult word detected from the input sentence, it is estimated whether it is used in the meaning of the high evaluation word or the low evaluation word from the nearby context, and the upper limit value is used as the word having the estimated meaning The information processing apparatus according to claim 1, wherein the information processing apparatus is used for setting a lower limit value.
The evaluation input restriction unit
The information processing apparatus according to claim 1, wherein information for specifying a high evaluation word and a low evaluation word in the input sentence is presented to the user.
The evaluation input restriction unit
The input sentence for enabling the evaluation value input with the value specified by the input operation when the input operation specifying the value outside the input permission range is performed as the evaluation value input operation The information processing apparatus according to claim 1, wherein information indicating a point requiring correction is presented to the user.
The evaluation input restriction unit
When an input operation for specifying a value outside the permitted input range is performed as an input operation for the evaluation value, another input sentence in which evaluation value input with the same value as the value specified in the input operation is permitted The information processing apparatus according to claim 1, which is presented to the user.
Accepting input of the input text and the evaluation value from the user for each item in a plurality of evaluation items, and accepting input of a comprehensive evaluation value separately from the evaluation value for each item,
The upper and lower limit value setting unit
Set an upper limit value and a lower limit value for the overall evaluation value according to the number of high evaluation words and low evaluation words in the entire input sentence across a plurality of the items,
The evaluation input restriction unit
Targeting the comprehensive evaluation value, presenting information that encourages modification of the input permission range and correction of the input text,
The information according to claim 1, further comprising: a wrinkle item notifying unit that determines whether or not there is a wrinkle between the input sentence and the input evaluation value for each item, and notifies the user of the item determined to have wrinkles. Processing equipment.
Upper / lower limit setting step for analyzing an input sentence by a user and setting an upper limit value and a lower limit value of an evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence;
When receiving the input of the evaluation value from the user, the input permission range of the evaluation value is limited to a range from the lower limit value to the upper limit value set by the upper and lower limit value setting step, and is outside the input permission range. An evaluation input restriction step of presenting information for prompting correction of the input sentence in response to an operation input instructing input of the evaluation value at
A correction rejection accepting step for accepting the denial operation of the correction;
The information processing apparatus executes a reference learning step of analyzing the input sentence in which the rejection operation is received by the correction rejection reception step, and correcting the upper limit value and the lower limit value setting criterion by the upper and lower limit value setting step. Information processing method.
Upper / lower limit value setting function for analyzing an input sentence by a user and setting an upper limit value and a lower limit value of an evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence;
When receiving the input of the evaluation value from the user, the input permission range of the evaluation value is limited to a range from the lower limit value to the upper limit value set by the upper and lower limit value setting function, and is outside the input permission range. An evaluation input restriction function for presenting information for prompting correction of the input sentence in response to an operation input instructing input of the evaluation value at
A correction rejection accepting function for accepting the denial operation of the correction;
Analyzing the input sentence in which the rejection operation is accepted by the correction rejection acceptance function, and causing the information processing apparatus to realize a reference learning function for correcting the upper limit value and the lower limit value setting criterion by the upper and lower limit value setting function program.
Upper / lower limit value setting function for analyzing an input sentence by a user and setting an upper limit value and a lower limit value of an evaluation value based on the number of high evaluation words and low evaluation words included in the input sentence;
When receiving the input of the evaluation value from the user, the input permission range of the evaluation value is limited to a range from the lower limit value to the upper limit value set by the upper and lower limit value setting function, and is outside the input permission range. An evaluation input restriction function for presenting information for prompting correction of the input sentence in response to an operation input instructing input of the evaluation value at
A correction rejection accepting function for accepting the denial operation of the correction;
Analyzing the input sentence in which the rejection operation is accepted by the correction rejection acceptance function, and causing the information processing apparatus to realize a reference learning function for correcting the upper limit value and the lower limit value setting criterion by the upper and lower limit value setting function A storage medium that stores programs.