JP2014182445A - Information processing device and navigation device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide information highly real-time related to a specific event without increasing a load on an information provider.SOLUTION: An information processing device includes: a comment collection part 1 for collecting a comment from a comment server publicizing submitted comments; a characteristics pattern database 3 for storing a characteristics pattern for detecting a keyword indicating time or period contained in the collected comment; a characteristics comment extraction part for referring to the characteristics pattern database 3 and extracting, as a characteristics comment, a comment containing time or period from the collected comment; a hot-word extraction part 4 for extracting, as a hot-word, a word with a higher frequency of appearance in the characteristics comment; a related information analysis part 5 for adding information, as related information, to a characteristic word of which the number of co-occurrence with a key word indicating time or period contained in the characteristics comment is equal to or higher than a threshold; and an information display part 6 for presenting with a characteristics comment containing the characteristics word added with the related information, together with the related information.

Description

  The present invention relates to an information processing apparatus and a navigation apparatus that acquire highly real-time information via a network and present it to a user.

In recent years, a plurality of manufacturers provide services that provide tourist information and event information to users using information processing devices such as navigation devices.
For example, in Patent Document 1, the peripheral information of a route to a designated point in a mobile object is acquired from an information providing server via an information network, and real-time information such as bargain information and event information is obtained for a user of the navigation system. An information display device that provides up-to-date information is disclosed.

JP 2004-257951 A

  However, in the technique disclosed in Patent Document 1 described above, in order to acquire the route peripheral information from the information providing server, the information providing server needs to hold detailed peripheral information over a wide range, In order to keep the information up-to-date, there is a problem that a great deal of cost is required.

  The present invention has been made to solve the above-described problems, and an object thereof is to provide highly real-time information regarding a specific event without increasing the burden on the information providing side.

  An information processing apparatus according to the present invention detects, from a comment server that publishes posted comments, a comment collection unit that collects comments, and a keyword that indicates a time or period included in the comments collected by the comment collection unit A feature pattern database that stores feature patterns, a feature comment extraction unit that refers to the feature pattern database, extracts comments including keywords indicating time or duration from comments collected by the comment collection unit, and a comment collection unit The collected comments and the feature comments extracted by the feature comment extraction unit are compared, and a feature word extraction unit that extracts words with higher appearance frequency as feature words in the feature comments, and the feature comments extracted by the feature comment extraction unit Key indicating the time or duration to be included A related information analysis unit that assigns information related to a keyword indicating time or period as related information to a feature word whose number of times of co-occurrence with a threshold is equal to or greater than a threshold value, and a related information analysis unit provided related information And an information display unit for presenting a feature comment including a feature word together with related information.

  According to this invention, it is possible to provide highly real-time information regarding a specific event without increasing the burden on the information providing side.

1 is a block diagram illustrating a configuration of an information processing device according to a first embodiment. It is a figure which shows an example of the feature pattern stored in the feature pattern database of the information processing apparatus by Embodiment 1. 3 is a flowchart illustrating an operation of the information processing apparatus according to the first embodiment. 6 is a diagram illustrating a display example of an information display unit of the information processing apparatus according to Embodiment 1. FIG. 6 is a block diagram illustrating a configuration of a navigation device according to Embodiment 2. FIG. 10 is a flowchart showing the operation of the navigation device according to the second embodiment. FIG. 10 is a block diagram illustrating a configuration of a navigation device according to a third embodiment. 12 is a flowchart illustrating an operation of the navigation device according to the third embodiment. FIG. 10 is a block diagram illustrating a configuration of a navigation device according to a fourth embodiment. 10 is a flowchart illustrating an operation at the time of route setting of the navigation device according to the fourth embodiment. 10 is a flowchart showing an operation when a route of the navigation device according to the fourth embodiment is moved.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing a configuration of an information processing apparatus according to Embodiment 1 of the present invention.
In FIG. 1, the information processing apparatus 10 includes a comment collection unit 1, a feature comment extraction unit 2, a feature pattern database 3, a hot word extraction unit (feature word extraction unit) 4, a related information analysis unit 5 and an information display unit 6. Has been.
The comment collection unit 1 accesses a comment server via a network, and collects comments posted by an unspecified number of users (hereinafter referred to as comments) and posting time information indicating the time when the comments are posted. . The collected comments and post time information are output as a comment group 101. An example of access to the comment server is a method of accessing a miniblog service such as twitter (registered trademark / hereinafter abbreviated) via the Internet. In the case of using twitter, an API that collects tweets posted within a predetermined period is disclosed (see Reference 1).
・ Reference 1
"Documentation | Twitter Developers", [online], [Searched on December 14, 2012], Internet <https://dev.twitter.com/docs>

  The feature comment extraction unit 2 refers to the feature pattern 102 stored in the feature pattern database 3 and extracts a feature comment from the comment group 101 collected by the comment collection unit 1. It is assumed that a time word is included in the feature comment to be extracted. Here, the time word is a word representing a relative date such as “today”, “next week”, “after N months”, or an absolute date such as “December 14” or “Autumn Day”. It is a word to represent. The feature comment extraction unit 2 outputs the extracted feature comments as the feature comment group 103 together with the comment group 101.

  The feature pattern database 3 is a database that stores a feature pattern 102 for the feature comment extraction unit 2 to detect a time word from the comment group 101. A time word is a word representing the relative and absolute dates described above. FIG. 2 shows an example of the feature pattern 102 stored in the feature pattern database 3. The feature pattern 102 includes “pattern information”, “pattern type”, “corresponding date”, and “date type”. Stores regular expressions representing time words such as “(. +) Month (. +) Day” as well as character strings indicating direct time words such as “today” and “next week” as pattern information for detecting the time word. To do. In the example of FIG. 2, a case where a character string and a regular expression are stored as pattern information is shown. However, the present invention is not limited thereto, and an automaton in which words are connected may be stored.

  The feature pattern database 3 also stores absolute or relative date information as the date corresponding to each stored pattern information. For example, relative dates of “± 0 day” and “+1 week” are stored in association with character strings “today” and “next week”, respectively. Further, the regular expression “(. +) Month (. +) Day” is stored in correspondence with the absolute date “$ 0 month $ 1 day”. However, $ 0 and $ 1 are numerical values that match the 0th and 1st regular expression groups, respectively.

  When a character string is stored in the feature pattern database 3, the feature comment extraction unit 2 determines whether or not each comment of the comment group 101 includes a stored character string. Extract. When the pattern information of the regular expression is stored in the feature pattern database 3, the feature comment extraction unit 2 searches for each comment using the regular expression and extracts it as a feature comment when a matching word exists. . Further, when pattern information other than character strings and regular expressions is stored in the feature pattern database 3, the feature comment extraction unit 2 determines the presence or absence of a time word in each comment by a method according to the pattern information, and the feature Extract comments.

  The hot word extraction unit 4 compares each comment in the comment group 101 input from the feature comment extraction unit 2 with each feature comment in the feature comment group 103, and selects a word having a higher appearance frequency in the feature comment as a hot word (feature Word). The hot word extraction unit 4 outputs the extracted hot word as the hot word group 104 together with the comment group 101. Here, the hot word is a keyword related to holding information of events that occur within a predetermined period such as the near future. Note that keywords related to various information can be extracted as hot words without being limited to events to be held.

  The related information analysis unit 5 extracts a comment including each hot word of the hot word group 104 from the comment group 101, analyzes the extracted comment, and assigns a word having a high probability of co-occurring with the hot word as the related information 105 To do. The hot word group 104 and the assigned related information 105 are output. The information display unit 6 includes a display screen for presenting information to the user, and presents the hot word group 104 and the related information 105 input from the related information analysis unit 5 via the display screen to the user. .

Next, the operation of the information processing apparatus 10 will be described.
FIG. 3 is a flowchart showing the operation of the information processing apparatus according to the first embodiment.
First, the comment collection unit 1 collects a comment posted by accessing the comment server and time information when the comment is posted (step ST1). The feature comment extraction unit 2 refers to the feature pattern database 3 and extracts feature comments including time words from the comments collected in step ST1 (step ST2).

The hot word extraction unit 4 decomposes the comments collected in step ST1 and the feature comments extracted in step ST2 into word units (step ST3).
In the decomposition of comments and feature comments into word units, for example, an open source morphological analysis engine “MeCab” disclosed in Reference Document 2 below is used.
・ Reference 2
MeCab: Yet Another Part-of-Speech and Morphological Analyzer, [online], [searched on December 14, 2012], Internet <http://mecab.googlecode.com/svn/trunk/mecab/doc/index. html>

  MeCab has a function of outputting parts of speech for each decomposed word. Since words related to events are often nouns including event names and facility names, calculation efficiency can be improved by limiting the words for which the appearance frequency is obtained to nouns. Note that it is not always necessary to use MeCab or other morphological analysis engines for decomposing comments and feature comments into words, and any means can be used as long as it has a function of analyzing sentences and decomposing them into words. May be used.

なお、式（１）においてｃｏｕｎｔ（ｗ＿Ｘｉ） はコメント群Ｘにおける単語ｗ＿Ｘｉの出現回数である。 Next, the hot word extraction unit 4 calculates the appearance frequency of each word of the comment and feature comment decomposed in step ST3 (step ST4).
When a set of comments is a comment group X (X is either a comment group 101 of comments collected by the comment collection unit 1 or a feature comment group 103 of feature comments extracted by the feature comment extraction unit 2). When the words appearing in the group X are w_X1, w_X2,..., W_Xn, the appearance frequency freq (X, w_Xi) of the certain word w_Xi (1 ≦ i ≦ n) in the comment group X uses the following equation (1). Is calculated.

In equation (1), count (w_Xi) is the number of appearances of the word w_Xi in the comment group X.

なお、式（２）においてｒａｎｋ（Ｘ， ｗ＿Ｘｉ（１≦ｉ≦ｎ））はコメント群Ｘにおける単語ｗ＿Ｘｉのランクを表す。 In addition, the method of calculating the appearance frequency of each word is not limited to the method using the above-described formula (1). It is good also as a method of using. However, the average rank value in a certain comment group is proportional to the word type n appearing in the comment group. Furthermore, since the word type n increases exponentially as the number of comments included in the comment group increases, it must be normalized as shown in the following equation (2) in order to be used for calculating the appearance frequency.

In Expression (2), rank (X, w_Xi (1 ≦ i ≦ n)) represents the rank of the word w_Xi in the comment group X.

  Next, the hot word extraction unit 4 uses the appearance frequency of each word calculated in step ST4 to generate the appearance frequency in the feature comment group 103 and the appearance frequency in the comment group 101 for each word included in the feature comment group 103. (Hereinafter referred to as an appearance frequency ratio r) is calculated (step ST5).

  The processing of step ST5 will be specifically described according to the above-described equation (1). Each word in the comment group 101 is set to w_A1, w_BA,..., W_Am, and each word in the feature comment group 103 is set to w_B1, w_B2,. In this case, for the words w_B1, w_B2,..., W_Bm in the feature comment group 103, the appearance frequency freq (B, w_Bk) in the feature comment group 103 and the appearance frequency ratio r of the appearance frequency freq (A, w_Bk) in the comment group 101. Is calculated. Note that k satisfies 1 ≦ k ≦ m. When the word w_Bk appears only in the feature comment group 103 (freq (A, w_Bk) = 0), r = 1.

  Thereafter, the hot word extraction unit 4 extracts, as hot words, words whose appearance frequency ratio r calculated in step ST5 exceeds a preset threshold Tr (step ST6). However, in the process of step ST6, the appearance frequency ratio r is usually increased with respect to the time word, so it is not included in the hot word.

  Next, the related information analysis unit 5 refers to the feature pattern 102 stored in the feature pattern database 3, analyzes all the time words included in the feature comment group 103 extracted in step ST2, and sets each time word. The corresponding absolute date is calculated (step ST7). In the process of step ST7, when the time word corresponds to a relative date, it is converted to an absolute date using the comment posting date. For example, when a certain time word corresponds to the relative date “+1 week” and the posting date and time of the comment including the time word is “Friday, December 14, 2012”, the related information analysis unit 5 Converts it to an absolute date and time indicating the Monday of the week following the posting date and time of “Monday, December 17, 2012”.

  Next, the related information analysis unit 5 analyzes the number of co-occurrence with the time word analyzed in step ST7 for the hot word extracted in step ST6 (step ST8). In the process of step ST8, when a hot word and a certain time word exist in the same comment, the number of co-occurrence of the hot word and the time word is increased by “1”.

  Next, with reference to the analysis result of step ST8, it is determined whether or not there is a hot word having a time word co-occurring a predetermined number of times (step ST9). If there is a hot word having a time word that co-occurs a predetermined number of times (step ST9; YES), the absolute date calculated in step ST7 is assigned to the hot word (step ST10). On the other hand, when there is no hot word having a time word co-occurring a predetermined number of times (step ST9; NO), the process is terminated.

The information display unit 6 selects a feature comment including a hot word to which an absolute date is assigned in step ST10 from the feature comment group 103, and associates it with the hot word (step ST11).
In the process of step ST11, when there are a plurality of feature comments including a hot word given an absolute date, the information display unit 6 calculates the information amount of each feature comment, and the feature comment having the largest calculated information amount is selected. Associate with hot word. For example, the number of unique hot words included in the feature comment is used as the information amount of the feature comment.

  Next, when presenting a feature comment to the user, a difference value between the absolute date given to each hot word and the current date is obtained, and hot words in which the obtained difference value exceeds a predetermined value are processed thereafter. (Step ST12). Next, the feature comment associated with the hot word not excluded from the processing target in step ST12 is presented to the user together with the absolute date given to the hot word in step ST10 (step ST13). End the process.

FIG. 4 is a diagram illustrating a display example of the information display unit of the information processing apparatus according to the first embodiment.
The display screen of the information display unit 6 shown in FIG. 4 includes an absolute date 62 such as “October 7 (Sunday)” and “October 7 (Sunday) 10 am to noon Display information 61 composed of characteristic comments 63 such as “A fish festival is held by a cooperative” is displayed.

  As described above, according to the first embodiment, the comment collecting unit 1 that accesses the comment server via the network and collects comments, and the feature that is a comment including a time word with reference to the feature pattern database 3 A feature comment extraction unit 2 that extracts comments, a hot word extraction unit 4 that compares the collected comments with the feature comments and extracts words having a higher appearance frequency as hot words, and co-occurs with the extracted hot words Generated within a predetermined period because it is provided with a related information analysis unit 5 that assigns a word having a high probability of being used as related information, and an information display unit 6 that presents the hot word and the related information together to the user Information related to events to be performed is acquired in real time via the network and presented to the user It can be.

  Moreover, according to this Embodiment 1, since it comprised so that the comment which many unspecified third parties contributed with respect to the comment server may be collected, and the information relevant to a hot word may be acquired, on the information provision side It is possible to provide highly real-time information without increasing the burden on a certain comment server.

Embodiment 2. FIG.
In the second embodiment, a navigation function is added to the functions of the information processing apparatus described in the first embodiment, and comments related to a specific area set on the navigation function side are collected and feature comments are extracted. The structure of the navigation apparatus which performs is shown.
FIG. 5 is a block diagram showing a configuration of the navigation device according to the second embodiment.
The navigation apparatus 20 according to the second embodiment is provided with a place name extraction unit 7 added to the function of the information processing apparatus 10 shown in FIG. As a navigation function, a route setting unit 21 and a map information database 22 are additionally provided. In the following description, the same or equivalent parts as those of the information processing apparatus 10 shown in the first embodiment are denoted by the same reference numerals as those used in FIG. 1 and description thereof is omitted or simplified.

  The route setting unit 21 receives an input from the user and calculates and sets an appropriate route from the current position of the user to the input destination. The set route is output to the place name extraction unit 7 as route information 201. The map information database 22 is a database that stores, as place name information 202, maps in various parts of the world, place names at each position on the map, information related to the place names, and the like. Here, the place names include names of facilities such as “Mt. Koya” and “Ghibli Museum” in addition to the names of prefectures and cities such as “Kyoto Prefecture” and “Nagoya City”.

  In addition, as the information related to the place name, for example, “the size of the area represented by the place name”, “the importance of the place name”, “the person name related to the place name”, and the like are stored. “Area of the area represented by the place name” is a numerical value from 1 to 5, and the larger the area, the smaller the value. Assign small values in order of country, prefecture, city, region, and facility name. The “importance of the place name” is a numerical value arbitrarily assigned by the creator of the map information database 22, and for example, a value from “0” to “9” is set. Set a larger value for the place name that the creator considers important. In the “person name related to the place name”, a name of a celebrity born in the prefecture or an active municipal mayor if the name is a municipality, or a name of a builder of the facility if a facility name is set. Note that the information related to the place name is not limited to the above-described information, and may be any information as long as it is information related to the name of each place.

  The place name extraction unit 7 refers to the route information 201 set by the route setting unit 21 and the place name information 202 stored in the map information database 22 and extracts place names existing on the set route. The extracted place names are output as the place name group 203. When the route set by the route setting unit 21 is long, the number of place names existing on the route becomes enormous. Therefore, the place name extracting unit 7 is set with the maximum number of extracted place names L_max in advance. When the number of place names existing on the route exceeds the maximum number of extracted place names L_max, the priority is obtained for each place name existing on the route, and the maximum number of place names L_max is extracted in descending order of priority. The priority of each place name is, for example, “area width represented by place name” × 10 + “place name” based on “area size represented by place name” and “importance of place name” stored in the map information database 22. "Importance of".

  The comment collection unit 1a accesses the comment server via the network, and posts the comment posted about each place name included in the place name group 203 extracted by the place name extraction unit 7 and the posting time information indicating the time when the comment is posted. collect. As access to the comment server, for example, in the case of access to twitter via the Internet, an API for obtaining a tweet including a specific keyword posted within a predetermined time has been released, and by using the API Only comments including place names can be collected. Moreover, the comment collection part 1a provides the name information specified as a keyword at the time of collection to each collected comment. For example, when comments relating to a place name “Mt. Koya” are collected using the API of tweet, place name information “Mt. Koya” is assigned to each comment. The collected comments, post time information, and place name information are output as a comment group 101a.

  The feature comment extraction unit 2a selects a comment related to one place name from the comment group 101a collected by the comment collection unit 1a, and refers to the feature pattern 102 stored in the feature pattern database 3 to comment on the selected one place name. Extract feature comments from. This process is performed for all comments in the comment group 101a. It should be noted that a time word is included in the feature comment extracted as in the first embodiment. The feature comment extraction unit 2a outputs the extracted feature comments as the feature comment group 103 together with the comment group 101a.

Next, the operation of the navigation device 20 according to the second embodiment will be described.
FIG. 6 is a flowchart showing the operation of the navigation device according to the second embodiment.
In the following, the same steps as those of the information processing apparatus according to the first embodiment are denoted by the same reference numerals as those used in FIG. 2, and the description thereof is omitted or simplified.
When the route setting unit 21 sets a route based on a user input (step ST21), the place name extraction unit 7 receives route information 201 input from the route setting unit 21 and place name information 202 stored in the map information database. , Place names existing on the set route are extracted (step ST22). The comment collection unit 1a accesses the comment server, and collects the comments posted for the names of each place extracted in step ST22 and the time information when the comments are posted (step ST23).

  The feature comment extraction unit 2a selects one place name from the place name information given to the comment collected in step ST23, and extracts a feature comment including a time word from the comment related to the selected place name with reference to the feature pattern database 3. (Step ST24). The hot word extraction unit 4 performs the above-described processing from step ST3 to step ST6 on the feature comment extracted in step ST24, thereby extracting hot words. Further, the related information analysis unit 5 performs the processing from step ST7 to step ST9.

  When it is determined in the determination process of step ST9 that there is a hot word having a time word that co-occurs a predetermined number of times (step ST9; YES), the absolute value corresponding to the time word is calculated using the date calculated in step ST7. Date and place name information is assigned to the hot word (step ST25). On the other hand, if it is determined in step ST9 that there is no hot word having a time word that co-occurs a predetermined number of times (step ST9; NO), the process proceeds to step ST26.

  The feature comment extraction unit 2a determines whether or not hot words have been extracted for all place names given to the comment (step ST26), and if hot words have not been extracted for all place names (step ST26) (ST26; NO), the process returns to step ST24. On the other hand, when hot words are extracted for all place names (step ST26; YES), the information display unit 6 performs the above-described processing of step ST11 and step ST12 based on the determination result, and is given to the hot words. Based on the place name information, the feature comment associated with each hot word and the absolute date assigned to the hot word are displayed on the map (step ST27), and the process ends.

  As described above, according to the second embodiment, the place name extracting unit 7 that extracts the place name existing on the route set by the user via the route setting unit 21 and the comment server via the network are accessed. Since it is configured to include a comment collecting unit 1a that collects comments and posting time information posted regarding the extracted names, and a feature comment extracting unit 2a that extracts feature comments from each collected comment, the user Information related to events that occur within a predetermined period in the vicinity of the route set by can be acquired in real time via the network and presented to the user.

  Moreover, according to this Embodiment 2, since it comprised so that the comment regarding the name of each place which the unspecified many third party posted with respect to the comment server may be collected, information relevant to a hot word will be acquired. It is possible to provide highly real-time information about the route set by the user without increasing the burden on the comment server on the providing side.

Embodiment 3 FIG.
In the third embodiment, in addition to the function of the navigation device 20 shown in the second embodiment, a configuration in which a feature comment and an absolute date are displayed on a map based on the current position of the user is shown.
FIG. 7 is a block diagram showing a configuration of the navigation device according to the third embodiment.
The navigation device 20a of the third embodiment is provided with a position information acquisition unit 23 added to the navigation device 20 shown in FIG. The information display unit 6a further includes a position information processing unit 6b as an internal configuration. In the following, the same or corresponding parts as the components of the navigation device 20 shown in the second embodiment are denoted by the same reference numerals as those used in FIG. 5 and the description thereof is omitted or simplified.

  The position information acquisition unit 23 always acquires information indicating the current position of the navigation device 20a using GPS information. Furthermore, when the navigation device 20a is mounted on the vehicle, the position information acquisition unit 23 always acquires the moving speed of the navigation device 20a using the measurement information of the vehicle speedometer. In addition, the structure which acquires the information which shows the present position of the navigation apparatus 20a, and a moving speed is not limited to what was mentioned above, A various application is possible if it is a means which can achieve the objective.

  The position information processing unit 6b of the information display unit 6a, based on the information indicating the current position of the navigation device 20a acquired by the position information acquisition unit 23, the position of the place name assigned to the hot word group 104, and the navigation device 20a. The position information processing unit 6b determines whether or not the distance from the current position is equal to or less than the threshold value T_dist based on the moving speed of the navigation device 20a acquired by the position information acquisition unit 23. Set T_dist. When the moving speed of the navigation device 20a is higher than the threshold value and faster, the threshold value T_dist is set larger, and when it is lower than the threshold value and slower, the threshold value T_dist is set lower. Thereby, when the moving speed of the navigation device 20a is fast, it is possible to prevent the user from passing the place name given to the hot word before the information presentation due to the delay of the information presentation by the information display unit 6a.

Next, the operation of the navigation device 20a will be described.
FIG. 8 is a flowchart showing the operation of the navigation device according to the third embodiment.
In the following, the same steps as those of the information processing apparatus according to the second embodiment are denoted by the same reference numerals as those used in FIG. 6, and the description thereof is omitted or simplified. In the flowchart of FIG. 8, steps ST21 to ST31 are operations at the time of route setting, and steps ST32 to ST36 are operations when moving on the route.

  When the processing from step ST21 to step ST12 shown in the second embodiment is performed and the information display unit 6a excludes hot words whose date difference value exceeds a predetermined value in step ST12, the map and the set route are displayed on the display screen. Is displayed (step ST31). Next, the position information processing unit 6b determines whether or not the route set by the user is moving based on the information indicating the current position of the navigation device 20a acquired by the position information acquisition unit and the moving speed (step). ST32). When the set route is not moved (step ST32; NO), the determination process of step ST32 is repeated.

  On the other hand, when moving along the set route (step ST32; YES), the position information processing unit 6b refers to the place name information given to each hot word in the hot word group 104 and the map information in the map information database 22. Then, the position indicated by the name of each place is calculated, and the distance from the current position of the navigation device 20a is calculated (step ST33). Further, the position information processing unit 6b determines whether or not the distance calculated in step ST33 is equal to or less than a threshold value T_dist (step ST34).

  If not less than or equal to the threshold T_dist (step ST34; NO), the process proceeds to step ST36. On the other hand, if the value is equal to or less than the threshold T_dist (step ST34; YES), the information display unit 6a displays the feature comment associated with each hot word on the map displayed on the display screen and the absolute value assigned to the hot word. The date is displayed (step ST35), and it is determined whether or not all place names have been processed (step ST36). If processing has not been performed for all place names (step ST36; NO), the processing returns to step ST33 and the above-described processing is repeated. On the other hand, when the process has been performed for all place names (step ST36; YES), the process ends.

  As described above, according to the third embodiment, the position information acquisition unit 23 that acquires the current position and moving speed of the navigation device 20a, and the acquired current position and moving speed of the navigation device 20a are used. Since the feature comment and the absolute date are displayed on the map when the distance between the position of each place name based on the place name information given to the hot word and the current position of the navigation device 20a is less than the threshold, It is possible to display information related to an event or the like in the vicinity of the route set by the user and according to the current position of the user.

  Further, according to the third embodiment, the information processing unit 6b is configured to set the threshold value T_dist based on the moving speed of the user, and thus displays information even when the moving speed of the user is fast. It is possible to prevent the user from passing through the corresponding place name before doing so.

Embodiment 4 FIG.
The fourth embodiment shows a configuration in which a user's utterance is recognized by speech and information based on the recognition result is presented.
FIG. 9 is a block diagram showing a configuration of the navigation device according to the fourth embodiment.
The navigation device 20b according to the fourth embodiment is provided with a recognition dictionary database 24, a sound collection unit 25, and a voice recognition unit 26 added to the navigation device 20 shown in FIG. The information display unit 6c is further provided with an audio processing unit 6d as an internal configuration. In the following, the same or corresponding parts as the components of the navigation device 20 shown in the second embodiment are denoted by the same reference numerals as those used in FIG. 5 and the description thereof is omitted or simplified.

  The voice processing unit 6d uses each hot word included in the hot word group 104, the place name assigned to each hot word, and information on the place name stored in the map information database 22 (for example, a person name related to the place name). The phonetic information 205 expressing the relationship between the word and the acoustic model is generated and stored in the recognition dictionary database 24.

  The recognition dictionary database 24 stores a speech pattern model obtained by modeling a time series pattern of acoustic feature vectors of a recognition vocabulary. In the fourth embodiment, the recognition dictionary database 24 describes an acoustic model in which features of speech are modeled, a syntax rule type language model in which connection rules between recognition units are defined, and a relationship between each recognition unit and the acoustic model. The generated pronunciation information 205 is stored. The acoustic model is an HMM (Hidden Markov Model), and the recognition unit in the language model is a word unit. In addition, when the pronunciation information 205 is added to the recognition dictionary database 24, if the number of acoustic models included in the word falls below a predetermined threshold, the addition to the recognition dictionary database 24 is not performed. As a result, it is possible to suppress adding words that are too short to the recognition dictionary database 24, and to suppress occurrence of unnecessary recognition results.

  The sound collection unit 25 always collects sound emitted in the vicinity of the navigation device 20b, and generates an audio signal 207 corresponding to the collected sound. The voice recognition unit 26 analyzes the voice signal 207 generated by the sound collection unit 25 and acquires an acoustic feature vector. In the fourth embodiment, the acoustic feature vectors are MFCC and ΔMFCC. Next, speech recognition is performed by the Viterbi algorithm using the stored information 206 of the recognition dictionary database 24. When the pronunciation information 205 corresponding to the audio signal 207 is stored in the recognition dictionary database 24, each hot word included in the hot word group 104, the place name assigned to each hot word, and the map information database 22 are stored. It can be determined that the information related to the place name is recognized as speech. A plurality of recognition score pairs indicating the likelihood of the recognized character string and recognition result are output as the speech recognition result 208. Note that the acoustic feature vectors used for speech recognition are not limited to MFCC and ΔMFCC, and for example, speech power or ΔΔMFCC may be used. The speech recognition algorithm is not limited to the Viterbi algorithm, and an algorithm such as WFST may be used.

  The information display unit 6c refers to the voice recognition result 208 of the voice recognition unit 26, and when one of the information related to each hot word, the place name assigned to each hot word, and the name of each place is recognized, The feature comment and absolute date are displayed above.

Next, the operation of the navigation device 20b will be described.
FIGS. 10 and 11 are flowcharts showing the operation of the navigation device according to the fourth embodiment. FIG. 10 shows the operation when performing route setting, and FIG. 11 shows the operation during route traveling.
First, the operation of the navigation device 20b when performing route setting will be described with reference to FIG.
When the information display unit 6c excludes hot words whose date difference value exceeds a predetermined value (step ST12), the voice processing unit 6d of the information display unit 6c displays each hot word, a place name assigned to each hot word, and map information. The phonetic information 205 expressing the relationship between the words and the acoustic model is generated from the information on the place names stored in the database 22 and stored in the recognition dictionary database 24 (step ST41), and the process is terminated.

Next, the operation of the navigation device 20b when moving on the set route will be described with reference to FIG. It is assumed that the information display unit 6c of the navigation device 20b displays a map and a set route on the display screen.
When the user utters, the sound collection unit 25 detects the utterance, collects the sound, and generates an audio signal 207 (step ST51). The speech recognition unit 26 analyzes the speech signal 207 generated in step ST51 and acquires an acoustic feature vector (step ST52). Furthermore, the speech recognition unit 26 performs speech recognition of the acoustic feature vector acquired in step ST52 using the acoustic model, language model, and pronunciation information that are the storage information 206 stored in the recognition dictionary database 24, and recognizes the recognized character string. , And a recognition score indicating the certainty of the result is output as a speech recognition result 208 (step ST53).

  The information display unit 6c refers to the speech recognition result 208 output in step ST53, and determines whether or not each hot word, the place name given to each hot word, and information related to the place name are included. Perform (step ST54). When the speech recognition result 208 does not include any of the hot word, the place name assigned to the hot word, or information on the place name (step ST54; NO), the process returns to step ST51 and the above-described process is repeated.

  On the other hand, when the speech recognition result 208 includes any one of the hot word, the place name assigned to the hot word, and information about the place name (step ST54; YES), the comment associated with the corresponding hot word, The absolute date assigned to the hot word is displayed on the map (step ST55), the process returns to step ST51, and the above-described process is repeated.

  As described above, according to the fourth embodiment, each hot word, the place name given to each hot word, and the place name stored in the map information database 22 are used to obtain the relationship between the word and the acoustic model. A speech processing unit 6d for generating the expressed pronunciation information 205, a speech recognition unit 26 for speech recognition of the speech signal 207 generated by the sound collection unit 25 to acquire an acoustic feature vector, and outputting a speech recognition result 208; When each of the hot words, the place name assigned to each hot word, and information related to each place name are recognized with reference to the recognition result 208, a feature comment and an absolute date are displayed on the map. Since the information display unit 6c is provided, when the user speaks with interest in a place name or event, the information display unit 6c is related to the place name or event. It can be presented clarity real-time information to the user.

  In the above-described fourth embodiment, the configuration in which the recognition dictionary database 24, the sound collection unit 25, the voice recognition unit 26, and the voice processing unit 6d are added to the configuration of the navigation device of the second embodiment is shown. You may apply to the information processing apparatus 10 of Embodiment 1 shown in FIG. 1, and the navigation apparatus 20b of Embodiment 3 shown in FIG.

  In the present invention, within the scope of the invention, any combination of the embodiments, or any modification of any component in each embodiment, or omission of any component in each embodiment is possible. .

  1, 1a comment collection unit, 2, 2a feature comment extraction unit, 3 feature pattern database, 4 hot word extraction unit, 5 related information analysis unit, 6, 6a, 6c information display unit, 6b position information processing unit, 6d voice processing Part, 7 place name extraction part, 10 information processing apparatus, 20, 20a, 20b navigation apparatus, 21 route setting part, 22 map information database, 23 position information acquisition part, 24 recognition dictionary database, 25 sound collection part, 26 speech recognition part .

Claims (14)

A comment collection unit that collects the comments from a comment server that publishes posted comments;
A feature pattern database for storing a feature pattern for detecting a keyword indicating a time or a period included in the comment collected by the comment collection unit;
A feature comment extraction unit that refers to the feature pattern database and extracts a comment including a keyword indicating the time or period as a feature comment from the comments collected by the comment collection unit;
A feature word extraction unit that compares the comments collected by the comment collection unit with the feature comments extracted by the feature comment extraction unit, and extracts words having a higher appearance frequency as feature words in the feature comments;
Information related to the keyword indicating the time or period is related information with respect to a feature word whose co-occurrence number with the keyword indicating the time or period included in the characteristic comment extracted by the characteristic comment extraction unit is equal to or more than a threshold. As a related information analysis unit
An information processing apparatus comprising: an information display unit that presents a feature comment including a feature word to which the related information analysis unit has added the related information together with the related information.   The related information analysis unit analyzes the keyword indicating the time or period included in the feature comment, calculates an absolute date and time that absolutely represents the date and time corresponding to the keyword, and calculates the calculated absolute date and time The information processing apparatus according to claim 1, wherein the information processing apparatus is provided as related information.   The information display unit calculates a difference value between the absolute date and time given as the related information and the current date and time, and when the calculated difference value exceeds a predetermined value, the feature word to which the related information is given The information processing apparatus according to claim 2, wherein a characteristic comment including a character string is not presented.   The information display unit calculates the information amount of each feature comment when there are a plurality of feature comments including the feature word to which the related information is given by the related information analysis unit, and the feature with the largest amount of information calculated The information processing apparatus according to claim 1, wherein a comment is selected and presented.   The information processing apparatus according to claim 4, wherein the information display unit calculates an information amount of the feature comment according to the number of the feature words included in each feature comment. A route setting unit for setting a route from the current position to the destination set by the user;
A map information database for storing map information, a place name at each position of the map information, and information related to the place name;
A place name extraction unit that extracts a place name existing on the route set by the route setting unit with reference to the map information database;
An information processing apparatus according to any one of claims 1 to 5,
The comment collecting unit collects comments including place names existing on the route set by the route setting unit from the comment server,
The information display unit refers to the map information database and displays a route set by the route setting unit, a feature comment including a feature word to which the related information is added, and the related information on a map. Navigation device.   The place name extraction unit is capable of setting a maximum number of place names indicating the number of place names to be extracted from the map information database, and refers to information related to the place names stored in the map information database. 7. The navigation apparatus according to claim 6, wherein a place name corresponding to the set maximum number of extracted place names is extracted based on a priority given to each place name existing on the route set by. A position information acquisition unit for acquiring the current position of the navigation device;
It is determined whether the distance between the current position of the navigation device acquired by the position information acquisition unit and the position of the place name given to any of the feature words extracted by the feature word extraction unit is less than or equal to a threshold value. A position information processing unit,
The information display unit, when it is determined that the distance between the current position of the navigation device and the position of any place name given to the feature word is equal to or less than a threshold value, the feature comment including the feature word The navigation apparatus according to claim 6, wherein the navigation apparatus is presented together with the related information. The position information acquisition unit acquires a current moving speed of the navigation device,
The position information processing unit determines a distance between the current position of the navigation device and the position of any place name given to the feature word according to the current moving speed of the navigation device acquired by the position information acquisition unit. The navigation apparatus according to claim 8, wherein a threshold value for determining whether or not is set. A speech processing unit that generates pronunciation information expressing a relationship between a word and an acoustic model based on a feature word to which the related information analysis unit has given the related information;
A recognition dictionary database for storing correspondence between pronunciation information and recognition vocabulary generated by the speech processing unit and acoustic feature vectors;
A sound collection unit that collects user utterances;
A speech recognition unit that recognizes the sound collected by the sound collection unit with reference to the recognition dictionary database;
The information display unit refers to a voice recognition result of the voice recognition unit, and when the pronunciation information corresponding to the voice collected by the sound collection unit is stored in the recognition dictionary database, the sound collection unit 6. The information processing apparatus according to claim 1, wherein a feature comment including a feature word corresponding to the collected voice is presented together with the related information.   The voice processing unit generates the pronunciation information for a feature word having a word length equal to or greater than a predetermined value among the feature words to which the related information analysis unit has added the related information. Information processing device. A speech processing unit that generates pronunciation information expressing a relationship between a word and an acoustic model based on a feature word to which the related information analysis unit has given the related information;
A recognition dictionary database for storing correspondence between pronunciation information and recognition vocabulary generated by the speech processing unit and acoustic feature vectors;
A sound collection unit that collects user utterances;
A speech recognition unit that recognizes the sound collected by the sound collection unit with reference to the recognition dictionary database;
The information display unit refers to a voice recognition result of the voice recognition unit, and when the pronunciation information corresponding to the voice collected by the sound collection unit is stored in the recognition dictionary database, the sound collection unit The navigation apparatus according to any one of claims 6 to 9, wherein a feature comment including a feature word corresponding to the collected voice is presented together with the related information.   The voice processing unit generates the pronunciation information for a feature word having a word length equal to or greater than a predetermined value among the feature words to which the related information analysis unit has added the related information. Navigation device.   The navigation device according to claim 12 or 13, wherein the voice processing unit generates pronunciation information based on the feature word, a place name given to the feature word, and information related to the place name. .

Images