CN113886541A

CN113886541A - Demand evaluation information generation method, demand evaluation information display method and device

Info

Publication number: CN113886541A
Application number: CN202111151198.XA
Authority: CN
Inventors: 李鸿宇; 万志文; 雷谦; 姚后清; 施鹏
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2021-09-29
Filing date: 2021-09-29
Publication date: 2022-01-04

Abstract

The disclosure provides a demand evaluation information generation method, a demand evaluation information display method and a demand evaluation information display device, and relates to the field of data processing, in particular to the field of retrieval statement processing. The specific implementation scheme is as follows: determining a target retrieval statement to be analyzed; acquiring retrieval amount information of a target retrieval statement in a first time period; the first time period is a time period before the current time, and the retrieval amount information of any retrieval statement in the time period comprises: the retrieval amount of the retrieval statement in each statistical cycle in the time period; determining an estimated quantity corresponding to the target retrieval statement based on the acquired retrieval quantity information; the pre-estimated quantity is used for representing the retrieval requirement of the target retrieval statement in a second time period, and the second time period is a time period after the current time; and generating the demand evaluation information of the target retrieval statement based on the pre-estimate amount corresponding to the target retrieval statement. According to the scheme, the demand evaluation information capable of accurately reflecting the actual demand can be generated.

Description

Demand evaluation information generation method, demand evaluation information display method and device

Technical Field

The disclosure relates to the technical field of data processing, in particular to the technical field of retrieval statement processing, and specifically relates to a method for generating demand evaluation information, a method for displaying demand evaluation information and a device thereof.

Background

Knowledge content production refers to the production of knowledge content meeting the requirements of users in multimedia forms of characters, pictures, videos, audios and the like in different fields. Since the search volume of the search statement (query) is an index that well reflects the user's needs, the commonly used strategy for knowledge content production is: knowledge contents corresponding to search sentences having a larger search amount are produced with higher priority.

In the related art, the average retrieval amount of a retrieval sentence in the last week or the total retrieval amount of the last three days is used as the demand evaluation information of the retrieval sentence, and the demand evaluation information is information indicating the production demand degree of the knowledge content.

Disclosure of Invention

On the basis, the present disclosure also provides a demand evaluation information display method and apparatus for displaying demand evaluation information capable of accurately reflecting actual demands.

According to an aspect of the present disclosure, there is provided a demand evaluation information generation method, including:

determining a target retrieval statement to be analyzed;

acquiring retrieval amount information of the target retrieval statement in a first time period; the first time period is a time period before the current time, and the retrieval amount information of any retrieval statement in a time period comprises: the retrieval amount of the retrieval statement in each statistical cycle in the time period;

determining an estimated quantity corresponding to the target retrieval statement based on the acquired retrieval quantity information; the pre-estimation is used for representing the retrieval requirement of the target retrieval statement in a second time period, wherein the second time period is a time period after the current time;

and generating the demand evaluation information of the target retrieval statement based on the corresponding pre-evaluation amount of the target retrieval statement.

According to another aspect of the present disclosure, a method for displaying demand evaluation information is provided, the method including:

acquiring screening information aiming at the retrieval statement;

screening target retrieval sentences matched with the screening information from the retrieval sentences stored in the requirement evaluation database; the demand evaluation database stores a plurality of retrieval statements and demand evaluation information of each retrieval statement, and the demand evaluation information of each retrieval statement is determined according to any one demand evaluation information generation method;

determining requirement evaluation information of the target retrieval statement from the requirement evaluation database; and outputting the target retrieval statement and the requirement evaluation information of the target retrieval statement.

According to another aspect of the present disclosure, there is provided a demand evaluation information generation apparatus including:

the first statement determination module is used for determining a target retrieval statement to be analyzed;

the first information acquisition module is used for acquiring retrieval amount information of the target retrieval statement in a first time period; the first time period is a time period before the current time, and the retrieval amount information of any retrieval statement in a time period comprises: the retrieval amount of the retrieval statement in each statistical cycle in the time period;

the estimated quantity determining module is used for determining estimated quantity corresponding to the target retrieval statement based on the acquired retrieval quantity information; the pre-estimation is used for representing the retrieval requirement of the target retrieval statement in a second time period, wherein the second time period is a time period after the current time;

and the information generation module is used for generating the demand evaluation information of the target retrieval statement based on the corresponding pre-estimation amount of the target retrieval statement.

According to another aspect of the present disclosure, there is provided a demand evaluation information presentation apparatus, the apparatus including:

the screening information acquisition module is used for acquiring screening information aiming at the retrieval statement;

the statement screening module is used for screening target retrieval statements matched with the screening information from the retrieval statements stored in the requirement evaluation database; the demand evaluation database stores a plurality of retrieval statements and demand evaluation information of each retrieval statement, and the demand evaluation information of each retrieval statement is determined by any one of the demand evaluation information generation devices;

the information determining module is used for determining the requirement evaluation information of the target retrieval statement from the requirement evaluation database;

and the information output module is used for outputting the target retrieval statement and the requirement evaluation information of the target retrieval statement.

According to another aspect of the present disclosure, there is provided an electronic device including:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor, the instructions being executable by the at least one processor to enable the at least one processor to perform a demand evaluation information generation method and a demand evaluation information presentation method.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute a demand evaluation information generation method and a demand evaluation information presentation method.

According to another aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements a demand evaluation information generation method and a demand evaluation information presentation method.

It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present disclosure, nor do they limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description.

Drawings

The drawings are included to provide a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is a schematic diagram according to a first embodiment of the present disclosure;

FIG. 2 is a schematic diagram according to a second embodiment of the present disclosure;

FIG. 3 is a schematic diagram according to a third embodiment of the present disclosure;

FIG. 4 is a schematic diagram according to a fourth embodiment of the present disclosure;

FIG. 5 is a schematic diagram according to a fifth embodiment of the present disclosure;

FIG. 6 is a schematic diagram according to a sixth embodiment of the present disclosure;

FIG. 7 is a block diagram of an electronic device for implementing a method for generating demand assessment information according to an embodiment of the present disclosure;

Detailed Description

Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which various details of the embodiments of the disclosure are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

Knowledge content production refers to producing knowledge content meeting user requirements in multimedia forms of characters, pictures, videos, audios and the like in different fields, for example: aiming at the client capable of realizing content search, business personnel can produce video, pictures, characters and other contents meeting the requirements of users, so that the contents meeting the requirements are provided for the users when the users search the contents by utilizing the client. Moreover, since the search volume of the search term (query) is an index that well reflects the user's needs, the commonly used strategy for knowledge content production is: knowledge contents corresponding to search sentences having a larger search amount are produced with higher priority.

With the development of the internet, a large number of non-repetitive search sentences are generated on the internet every day, wherein the high-frequency search sentences account for about 20%, the medium-long-tailed search sentences account for more than 40%, the high-frequency search sentences are search sentences with a weekly average search quantity of more than 10, and the medium-long-tailed search sentences are search sentences with a weekly average search quantity of more than 2 and not more than 10.

How to reasonably quantify the value of the retrieval statement under limited productivity so as to efficiently utilize the retrieval statement and guide a knowledge service line to generate knowledge content capable of meeting the requirements of users is a core challenge faced by the current industry.

In the related art, the average retrieval amount of a retrieval sentence in the last week or the total retrieval amount of the last three days is used as the requirement evaluation information of the retrieval sentence. However, the solution of the related art does not take into consideration the problem that the search amount of most search terms changes significantly with time. Thus, for the time-efficient hot-point retrieval statement, if the production period is long, not only the knowledge content meeting the user requirement cannot be generated immediately, but also the produced knowledge content has poor time efficiency, so that the utilization degree cannot be expected, and the input-output ratio is low.

It can be seen that the related art has at least the following technical problems:

1. since the retrieval amount of the retrieval statement changes obviously with time, if the average retrieval amount in the last week or the retrieval total amount of the last three days is taken as the requirement retrieval information, the requirement retrieval information cannot accurately reflect the actual requirement;

2. knowledge content matching with actual requirements often cannot be generated by using requirement retrieval information provided by related technologies.

In order to solve the problem that the requirement evaluation information cannot accurately reflect the actual requirement in the related art, the embodiment of the disclosure provides a method for generating the requirement evaluation information.

It should be noted that, in a specific application, the requirement evaluation information generating method provided by the embodiment of the disclosure may be applied to various electronic devices, for example, a personal computer, a server, and other devices with data processing capability. In addition, it is understood that the requirement evaluation information generation can be realized by software, hardware or a combination of software and hardware. Moreover, the demand evaluation information generation method can be applied to any scene with demand evaluation, such as: for a client capable of realizing content search, a user can perform content retrieval through the client, and then, for a retrieval statement input when the user performs content retrieval, demand evaluation can be performed through the demand evaluation information generation method; or, for a client capable of playing multimedia content, a user may search for the multimedia content through the client, and then, for a search statement input when the user searches for the multimedia content, the requirement evaluation may be performed through the requirement evaluation information generation manner.

The method for generating demand evaluation information provided by the embodiment of the present disclosure may include:

determining a target retrieval statement to be analyzed;

acquiring retrieval amount information of a target retrieval statement in a first time period; the first time period is a time period before the current time, and the retrieval amount information of any retrieval statement in the time period comprises: the retrieval amount of the retrieval statement in each statistical cycle in the time period;

determining an estimated quantity corresponding to the target retrieval statement based on the acquired retrieval quantity information; the pre-estimated quantity is used for representing the retrieval requirement of the target retrieval statement in a second time period, and the second time period is a time period after the current time;

and generating the demand evaluation information of the target retrieval statement based on the pre-estimate amount corresponding to the target retrieval statement.

According to the scheme provided by the disclosure, the estimated quantity of the target retrieval statement in the second time period after the current time can be determined by utilizing the retrieval quantity information of the target retrieval statement in the first time period before the current time, and then the demand evaluation information of the target retrieval statement is generated based on the determined estimated quantity. Since the prediction quantity characterizes the retrieval demand of the target retrieval statement in the second time period, the demand evaluation information can accurately reflect the actual demand of the target retrieval statement in the future second time period. Therefore, the scheme provided by the disclosure can solve the problem that the actual requirements cannot be accurately reflected by the requirement evaluation information in the related technology.

Furthermore, because the requirement evaluation information can accurately reflect the actual requirement, when the requirement evaluation information is used for guiding the production of the knowledge content, the knowledge content matched with the actual requirement can be produced, and the scheme provided by the disclosure provides a realization basis for producing the knowledge content matched with the actual requirement.

A method for generating demand evaluation information according to an embodiment of the present disclosure is described below with reference to the drawings.

As shown in fig. 1, a method for generating demand evaluation information provided by the embodiment of the present disclosure may include the following steps:

s101, determining a target retrieval statement to be analyzed;

the target retrieval statement is a retrieval statement needing to determine the requirement evaluation information. The target retrieval statement may be determined in different manners in combination with the actual application scenario, for example, the manner of determining the target retrieval statement includes at least one of the following two manners:

in the first determination method, when there is a need to analyze a specified search term, the search term specified by the received specifying operation may be used as the target search term. The designation operation may include a selection operation, an input operation, and the like.

In the second determination method, when there is a need to analyze a search term that meets the filtering condition, the search term that meets the filtering condition may be used as the target search term. The above-mentioned screening condition may be determined according to requirements and experience, for example, belonging to a certain time period, or a retrieval frequency higher than a predetermined frequency threshold within a certain time period, etc., which may be a time period before a time when there is a requirement for statement analysis.

In addition, in one implementation, the screening condition may be: if the retrieval frequency of the retrieval statement in the first time period is greater than the preset frequency threshold, then determining the target retrieval statement to be analyzed in this step may include:

and screening the retrieval sentences of which the retrieval frequency is greater than a preset frequency threshold value from the retrieval sentences in the first time period to serve as target retrieval sentences to be analyzed. By screening the target retrieval sentences with the retrieval frequency greater than the preset frequency threshold, the target retrieval sentences can be ensured to be retrieval sentences with certain requirements.

The preset frequency threshold may be determined according to requirements and experience, for example, 1 time/day, 2 times/week, etc.

The first time period may be a time period before the current time, wherein the current time is a time for executing the method. Optionally, in an implementation manner, the first time period may be a time period with the current time as an end time and the duration as a first specified duration, where the first specified duration may be determined according to requirements and experience, for example, 3 days, 1 week, 10 days, and the like, and if the first specified duration is 1 week and the current time is 2021 year, 9 month, 8 days, the first time period is [2021 year, 9 month, 2 days, 2021 year, 9 month, 8 days ]. Alternatively, in another implementation, the first time period may be N1 consecutive statistical periods before the period of the current time, e.g., the statistical period is one week, and then the first time period is N1 consecutive weeks before the week of the current time, i.e., N1 consecutive weeks before the current week.

In the above process of screening the search sentences of which the search frequency is greater than the preset frequency threshold from the search sentences in the first time period, the search sentences in the first time period may be obtained first, and then the number of times of occurrence of each search sentence in the first time period may be counted. And then, the ratio of the occurrence frequency of each retrieval statement to the first time period can be calculated to obtain the retrieval frequency of each retrieval statement, and then the retrieval statements with the retrieval frequency larger than the preset frequency threshold are screened out according to the retrieval frequency of each retrieval statement.

In addition, it can be understood that, when the target retrieval statement is selected, the selected time period is the first time period, which is the time period used for acquiring the retrieval amount information of the target retrieval statement in the subsequent S102, so that it can be ensured that the target retrieval statement is the statement appearing in the first time period, and the problem of data invalidation caused by the fact that the target retrieval statement does not appear in the first time period after the target retrieval statement is screened in other time periods is avoided.

S102, acquiring retrieval amount information of a target retrieval statement in a first time period;

the retrieval amount information of any retrieval statement in a time period comprises: the retrieval amount of the retrieval statement in each statistical cycle in the time period. Each statistical period may be determined empirically and on demand, and may be, for example, 1 day or 1 week, etc.

After the target search statement is determined, the search amount of the target search statement in each statistical cycle in the first time period may be read from the search data generated in the first time period.

Illustratively, as shown in table 1, the retrieval amount distribution of the target retrieval statement in each statistical cycle in the first time period is obtained.

TABLE 1

Statistical period 1	Statistical period 2	Statistical period 3	Statistical period 4
				3	4	2	1

The retrieval amount information of the target retrieval statement may include (3,4,2,1), or include { (statistical period 1,3), (statistical period 2,4), (statistical period 3,2), (statistical period 4,1) }. S103, determining an estimated quantity corresponding to the target retrieval statement based on the acquired retrieval quantity information;

the estimate is used to represent the retrieval requirement of the target retrieval statement in the second time period, for example, the estimate may be the total retrieval quantity of the target retrieval statement in the second time period, or the estimate may also be the retrieval quantity of the target retrieval statement in each statistical cycle of the second time period, or the estimate may also be the average retrieval quantity of the target retrieval statement in each statistical cycle of the second time period, and so on.

Since the retrieval amount information includes the retrieval amount of the target retrieval statement in each statistical cycle in the first time period, the retrieval amount information of the target retrieval statement can reflect the variation trend of the retrieval amount of the target retrieval statement, so that the estimated amount of the target retrieval statement in the second time period can be predicted based on the variation trend of the retrieval amount of the target retrieval statement.

For example, the retrieval amount of the target retrieval statement in each statistical cycle in the first time period may be fitted to obtain a target change function of time and the retrieval amount, and then the estimated amount of the target retrieval statement in the second time period may be determined by using the target change function. Alternatively, the estimated amount of the target search statement in the second time period may also be determined in a neural network model manner, which will be specifically described in detail in the following embodiments and will not be described herein again.

In addition, the second time period is a time period after the current time. Optionally, in an implementation, the second time period may be a time period with the current time as a starting time and the duration of the second time period is a second designated duration, where the second designated duration may be determined according to requirements and experience, such as 3 days, 1 week, 10 days, etc., and if the second designated duration is 1 week, when the time is 2021 year, 9 month, 8 days, then the second time period is [2021 year, 9 month, 8 days, 2021 year, 9 month, 15 days ]. Alternatively, in another implementation, the second time period may be N2 consecutive periods after the period of the current time, e.g., one week, then the second time period is N2 consecutive weeks after the week of the current time, i.e., N2 consecutive weeks after the week.

It is emphasized that, in order to improve the accuracy of the determined estimate, the duration of the first time period may not be less than the duration of the second time period, for example, the first time period may be equal to or greater than the duration of the second time period.

And S104, generating the demand evaluation information of the target retrieval statement based on the pre-estimate amount corresponding to the target retrieval statement.

In this step, the estimated quantity of the target retrieval statement in the second time period is obtained, and the requirement evaluation information of the target retrieval statement can be further determined. There are many ways to generate the demand evaluation information, and examples include at least one of the following:

the first mode is that the obtained estimated quantity is used as the demand evaluation information of the target retrieval statement;

and in the second mode, the requirement level corresponding to the obtained estimated quantity is determined based on the corresponding relation between the pre-constructed estimated quantity and the requirement level and is used as the requirement evaluation information of the target retrieval statement.

The requirement level can be determined according to requirements and experiences, such as high requirements, medium requirements and low requirements, or primary requirements, secondary requirements, tertiary requirements and the like. The corresponding relation between the estimated quantity and the demand level can be divided according to the demand and experience, each demand level can correspond to an estimated quantity range, and the estimated quantities belonging to the estimated quantity range all correspond to the demand level.

According to the scheme provided by the disclosure, the retrieval amount information of the target retrieval statement in the first time period before the current time can be utilized to determine the estimated amount of the target retrieval statement in the second time period after the current time, and then the demand evaluation information of the target retrieval statement is generated based on the determined estimated amount. Therefore, the scheme provided by the disclosure can solve the problem that the actual requirements cannot be accurately reflected by the requirement evaluation information in the related technology.

Optionally, in another embodiment of the present disclosure, after generating the requirement evaluation information of the target search statement based on the pre-estimate corresponding to the target search statement, the generating manner of the requirement evaluation information may further include:

and writing the target retrieval statement and the requirement evaluation information of the target retrieval statement into a requirement evaluation database.

It can be understood that, when a plurality of target retrieval statements are provided, when a plurality of target retrieval statements and the requirement evaluation information of each target retrieval statement are written into the requirement evaluation database, clustering can be performed on each target retrieval statement to obtain a cluster title and the requirement evaluation information corresponding to the cluster title; and correspondingly recording cluster titles and the requirement evaluation information, and simultaneously recording the hierarchical relationship between the cluster titles and the corresponding target retrieval sentences. Therefore, when the demand evaluation database is subsequently inquired, the cluster title and the corresponding demand evaluation information can be preferentially displayed, and when an instruction for further detailed display is obtained, each target retrieval statement corresponding to the cluster title and the corresponding demand platform information are displayed. By the method, information recording and output can be more hierarchical.

In addition, the target retrieval statement and the demand evaluation information of the target retrieval statement can be written into the demand evaluation database periodically, specifically, the current date can be selected periodically every week, each target retrieval statement to be predicted is determined, the demand evaluation information of the target retrieval statement is determined by adopting the demand evaluation information generation method provided by the embodiment of the disclosure, and the demand evaluation information of the target retrieval statement and the target retrieval statement is written into the demand evaluation database. Therefore, the requirement evaluation database can be ensured to store the newer target retrieval statement and the requirement evaluation information of the target retrieval statement in real time.

According to the scheme provided by the disclosure, the demand evaluation information capable of accurately reflecting the actual demand can be generated, and a realization basis is provided for producing the knowledge content matched with the actual demand. Furthermore, the target retrieval statement and the requirement evaluation information of the target retrieval statement are written into the requirement evaluation database, so that a realization basis can be provided for subsequently utilizing the target retrieval statement and the requirement evaluation information of the target retrieval statement.

Based on the embodiment shown in fig. 1, as shown in fig. 2, a method for generating demand evaluation information according to another embodiment of the present disclosure may include the following steps:

s201, determining a target retrieval statement to be analyzed;

s202, acquiring retrieval amount information of a target retrieval statement in a first time period;

the first time period is a time period before the current time, and the retrieval amount information of any retrieval statement in the time period comprises: the retrieval amount of the retrieval statement in each statistical cycle in the time period;

in this embodiment, S201 to S202 are the same as S101 to S102 in the above embodiment, and are not described herein again.

S203, constructing characteristic data of the target retrieval statement in a first time period based on the acquired retrieval amount information;

in this step, the feature data may be a feature vector, and at this time, the retrieval amount of the target retrieval statement in each statistical period in the first time period may be obtained from the retrieval amount information, and then the retrieval amount of the target retrieval statement in each statistical period is used as a dimension value of the feature vector.

Illustratively, the retrieval amount information is: { (statistic period 1,3), (statistic period 2,4), (statistic period 3,2), (statistic period 4,1) }, the constructed feature vector is (3,4,2, 1).

Optionally, the feature data may further include another feature value for processing the search amount information, and in this case, the constructing the feature data of the target search term in the first time period based on the acquired search amount information may include:

and calculating the mean and/or variance of each retrieval amount in the acquired retrieval amount information, and constructing a feature vector as feature data of the target retrieval statement in the first time period by using the retrieval amounts included in the acquired retrieval amount information and the calculated mean and/or variance.

In this step, if the average value of each retrieval amount in the acquired retrieval information needs to be calculated, the sum of each retrieval amount may be calculated first, and then the calculated sum of the retrieval amounts is divided by the number of the statistical periods in the first time period, so as to obtain the average value of the target retrieval statement in each statistical period in the first time period.

If the variance of each search amount in the acquired search information is calculated, each search amount can be substituted into a variance calculation formula to obtain the variance of each search amount.

After the mean/variance of each retrieval amount is calculated, a feature vector may be constructed based on the retrieval amount included in the acquired retrieval amount information and the calculated mean/variance. In one implementation, the respective search variables and the calculated mean and/or variance may be directly used as data of one dimension of the feature vector. In another implementation, each search quantity and the calculated mean/variance may be subjected to a log (log) transformation, and the logarithmically transformed values may be used as data of one dimension of the feature vector.

S204, processing the characteristic data of the target retrieval statement by using a pre-trained predictor prediction model to obtain a predictor corresponding to the target retrieval statement;

wherein, the prediction model of the pre-estimation amount is as follows: the method comprises the steps that training is carried out on the basis of sample data and labeled data corresponding to the sample data, wherein the sample data is characteristic data of a specified retrieval statement in a sample time period, the labeled data is used for representing retrieval requirements of the specified retrieval statement in a labeled time period, the sample time period is a time period before reference time, and the labeled time period is a time period after the reference time.

The reference time may be any specified historical time, and for example, the reference time is 1/15/2021, the sample time period may be [ 1/8/2021, 1/15/2021 ], and the labeled time periods are [ 1/15/2021, 1/22/2021 ].

Optionally, the duration of the sample time period may be the same as the duration of the first time period, and the duration of the labeled time period is the same as the duration of the second time period. Further, the time interval between the sample time period and the labeled time period is the same as the interval between the first time period and the second time period, and is 0, for example.

Specifically, the feature data of the target search term may be input to the predictor prediction model, and the data output by the predictor prediction model may be used as the predictor corresponding to the target search term.

The marked data corresponding to the sample data may be the total retrieval amount of the specified retrieval statement in the marked time period, or the estimated amount may also be the retrieval amount of the specified retrieval statement in each statistical period of the marked time period, or the estimated amount may also be the average retrieval amount of the specified retrieval statement in each statistical period of the marked time period, which is reasonable.

The specific training mode of the prediction model will be described in the following embodiments, and will not be described herein.

And S205, generating the demand evaluation information of the target retrieval statement based on the pre-estimate amount corresponding to the target retrieval statement.

In this embodiment, S205 is the same as S104 in the above embodiment, and is not described herein.

According to the scheme provided by the disclosure, the demand evaluation information capable of accurately reflecting the actual demand can be generated, and a realization basis is provided for producing the knowledge content matched with the actual demand. Furthermore, the estimated quantity of the target retrieval statement in the second time period can be predicted by using the estimated quantity prediction model, so that a realization basis is provided for generating the demand evaluation information capable of accurately reflecting the actual demand, and the estimated quantity of the target retrieval statement in the second time period can be accurately and quickly predicted by using the estimated quantity detection model.

Based on the embodiment shown in fig. 2, as shown in fig. 3, in the method for generating demand evaluation information according to another embodiment of the present disclosure, a training mode for training a predictive model is as follows:

s301, acquiring first retrieval amount information of a specified retrieval statement in a sample time period and second retrieval amount information of the specified retrieval statement in a labeling time period from a training data set;

the training data set may include first retrieval amount information of each retrieval statement in a sample time period and second retrieval amount information of each retrieval statement in a labeling time period. The specified search statement may be any search statement in a sample time period, or a search statement with a search frequency greater than a preset frequency threshold in the sample time period, and after the search statement is determined, first search amount information of the specified search statement in the sample time period and second search amount information of the specified search statement in a labeling time period may be acquired from the training data set.

The training data set may be pre-constructed. Optionally, a reference time may be pre-specified, and then a sample time period after the reference time and a labeling time period before the reference time are determined, so as to construct a training data set.

In an implementation manner, after a sample time period and a labeling time period are determined, each retrieval statement existing in the sample time period can be determined, first retrieval amount information of each retrieval statement in the sample time period and second retrieval amount information of each retrieval statement in the labeling time period are further obtained, and finally, a training data set is constructed by using each retrieval statement and the first retrieval amount information and the second retrieval amount information of each retrieval statement.

It is emphasized that the training data set may include a plurality of pairs of sample time periods and labeled time periods, and each pair of sample time period and labeled time period corresponds to a reference time. As shown in table 2:

TABLE 2

Training data set	Sample time period	Annotating time periods	Reference time
				Training set 1	20200422	20200930	20200708
Training set 2	20200429	20201007	20200715
				Training set 3	20200506	20201014	20200722
…	…	…	…
				Training set n	20201111	20210421	20210127

Each training set is a pair of a sample time period and a labeling time period, the training set 1 is taken as an example, the reference time is 07/08/2020, the sample time period is 22/04/2020, 07/08/2020, and the labeling time period is 30/09/2020.

S302, generating characteristic data of the specified retrieval statement in a sample time period as sample data based on the first retrieval amount information, and generating labeled data of the specified retrieval statement in a labeled time period based on the second retrieval amount information;

the process of determining the feature data of the specified search statement in the sample time period is the same as or similar to the process of determining the feature data of the target search statement in the first time period, and is not repeated here.

The marking data of the specified retrieval statement in the marking time period may include: and marking at least one of the total retrieval amount in the time period, the retrieval amount in each statistical period in the time period, the average retrieval amount in each statistical period in the time period, and the like.

If the tagged data includes the total search amount, the search amounts in each statistical period in the second search amount information may be added to obtain the total search amount. If the annotation data includes the retrieval amount in each statistical cycle in the annotation time period, the second retrieval amount information may be directly used as a part of the annotation data. If the annotation data includes the average retrieval amount in each statistical period in the annotation time period, the retrieval amount in each statistical period in the annotation time period can be obtained after the retrieval total amount is calculated by using the second retrieval amount information and the retrieval total amount is averaged.

Optionally, the annotation data may be determined according to requirements, and its form may be various, for example:

in one mode, only the total retrieval amount in the labeling time period or the average retrieval amount in each statistical period in the labeling time period can be used as labeling data;

alternatively, the annotation data may include a plurality of values, such as a retrieval amount in each statistical period in the annotation time period, or at least two of the retrieval amount in each statistical period in the annotation time period, a total retrieval amount in the annotation time period, and an average retrieval amount in each statistical period in the annotation time period.

For example, in a first scenario, if the annotation data is a retrieval amount in each statistical cycle in the annotation time period, the annotation data may be (a1, a 2.., An), where a1, a 2.., An is a retrieval amount of the specified retrieval statement in each statistical cycle in the annotation time period;

in a second scenario, the annotation data comprises: marking the retrieval amount in each statistical period in the time period and the retrieval total amount in the marking time period, wherein B is the retrieval total amount in the marking time period, and the marking data can be (a1, a 2., An, B);

in a third scenario, the annotation data comprises: marking the retrieval amount in each statistical period in the time period and the average retrieval amount in each statistical period in the time period, wherein B is the average retrieval amount in each statistical period in the marking time period, and the marking data can be (a1, a2,. ·, An, C);

in a fourth scenario, the annotation data comprises: and (4) marking the total retrieval amount in the time period and the average retrieval amount in each statistical period in the time period, so that the marking data can be (B, C).

S303, inputting sample data into the neural network model to be trained, so that the neural network model predicts the pre-estimated quantity of the specified retrieval statement in the labeling time period based on the sample data and takes the pre-estimated quantity as prediction data;

in order to train the neural network model to be trained, in this step, sample data may be input to the neural network model to obtain an estimated value of the specified sample model predicted by the neural network model in the labeled time period.

S304, calculating a loss function value of the neural network model based on the prediction data and the labeling data;

after the three-dimensional data output by the neural network model is obtained, a loss function value of the neural network model can be calculated based on the prediction data and the marking data, so that the difference between the prediction data output by the neural network model and the ideal marking data can be represented through the loss function value.

Optionally, in an implementation manner, a difference value between the prediction data and the same dimension data as the annotation data may be calculated, and a sum of the difference values in each dimension is calculated to serve as a loss function value of the neural network model.

Illustratively, if the annotation data is (a1, a 2.., an), and the prediction data is (b1, b 2.., bn), then

S305, adjusting parameters of the neural network model according to the loss function value;

for the neural network model, the larger the loss is, the larger the adjustment range of the parameter to be adjusted is, and therefore, the parameters of the neural network model can be adjusted based on the loss result by combining the actual situation and the requirement. Optionally, the neural network model parameters may be adjusted based on the loss function values through a gradient descent equal parameter adjustment algorithm.

S306, judging whether the search amount information in the training sample set is used, if so, ending the training, otherwise, returning to execute S301.

After the neural network model parameters are adjusted according to the loss function values, the next training can be carried out until the retrieval amount information in the training sample set is utilized.

According to the scheme, the demand evaluation information capable of accurately reflecting the actual demand can be generated, a realization basis is provided for producing knowledge content matched with the actual demand, and further, a realization basis is provided for generating the demand evaluation information capable of accurately reflecting the actual demand through the training pre-estimate model.

It should be noted that the prediction model in this embodiment is not a model for a specific user, and cannot reflect personal information of a specific user.

In the technical scheme of the present disclosure, the processes of collecting, storing, using, processing, transmitting, providing, disclosing and the like of the personal information of the related users all conform to the regulations of the related laws and regulations, and do not violate the good custom of the public order. The retrieval statement in this embodiment may be from a public data set.

In order to better understand the scheme provided by the present disclosure, in an embodiment of the present disclosure, taking 12 weeks as an example of a specified duration, the following scheme provided by the present disclosure is introduced:

first, a target retrieval statement may be determined based on log data searched by the mobile end user in the current week (i.e., in the current week), or the target retrieval statement may be specified, and the retrieval amount of the target retrieval statement in the history of 12 weeks between the weeks may be acquired.

Then, the search volume of the target search term in 12 weeks of the history and the mean and variance of the search volume of the target search term in 12 weeks are logarithmically transformed to obtain transformed values as feature data of the target search term.

And finally, inputting the obtained characteristic data into a pre-constructed prediction quantity prediction model to obtain the sum of the search quantity of the current future 12-week target search statement.

The prediction model can be obtained by training based on the retrieval amount of the historical retrieval sentences.

For example, when the sunday 2021 year 5 month 5 day, search data of search sentences from 22/month 4/2020 to 21/month 4/2021 year can be acquired as training set data as shown in table 3, and a plurality of training sets are divided at different reference times as shown in table 3:

TABLE 3

Training data set	Sample time period	Annotating time periods	Reference time
				Training set 1	20200422	20200930	20200708
Training set 2	20200429	20201007	20200715
				Training set 3	20200506	20201014	20200722
…	…	…	…
				Training set n	20201111	20210421	20210127
Test set	20201118	20210428	20210203

Further, test data, such as test set data with 2021 year 2 month 3 day as a reference time in table 3, may be added additionally.

It should be noted that each date in table 3 is wednesday in the week, i.e., each wednesday represents the data of the week.

After the training data and the test data are obtained, the prediction quantity model can be trained by using the training data, and the test data is used for testing the prediction quantity model after being trained correspondingly until the prediction quantity model meeting the requirements is obtained.

After the sum of the retrieval amounts of the target retrieval statement in the future 12 weeks is obtained through the estimated quantity prediction model, the sum of the retrieval amounts of the target retrieval statement in the historical 12 weeks and the sum of the retrieval amounts of the target retrieval statement in the future 12 weeks can be stored in an ElasticSearch database for data storage, wherein the ElasticSearch database is a database with a real-time engine for distributed storage, search and analysis.

As shown in fig. 4, the method for displaying demand evaluation information provided in the embodiments of the present disclosure may include the following steps:

s401, obtaining screening information aiming at a retrieval statement;

the screening information may be determined according to the received screening operation of the user, and the screening information is the screening condition. Optionally, the front-end page may be displayed in advance, and the front-end page may display different information to the user as required, such as information of a cluster title option, a PV (day-of-week average retrieval) range, a keyword, a demand evaluation information generation date, a type of knowledge content production, and the like. Optionally, the information may be visually presented.

Of course, other information may also be present in the front page to provide richer options for information screening. For example: and (3) classifying the requirements: the method comprises the following steps of first-level classification and second-level classification, wherein the first-level classification and the second-level classification are field classification, and the second-level classification is refined classification of the first-level classification. When receiving operations such as selection and input on the front page by a user, filtering information for the search statement can be generated according to the operations of the user. The present disclosure does not limit the screening information provided.

S402, screening target retrieval sentences matched with screening information from the retrieval sentences stored in the requirement evaluation database;

the demand evaluation database stores a plurality of retrieval statements and demand evaluation information of each retrieval statement, and the demand evaluation information of each retrieval statement is determined according to the demand evaluation information generation method provided by the disclosure;

the determination method of the requirement evaluation information of each search statement is described in the foregoing embodiments, and is not described herein again. After the filtering information is obtained, the target retrieval statement matched with the filtering information can be filtered from the retrieval statements stored in the requirement evaluation database.

For example, the screening information is a keyword for screening the education industry and including "middle school", and then target retrieval sentences such as "how to learn in the middle school language" and "a learning method of physics in the middle school" can be screened out.

It is understood that the requirement evaluation database may also record description information of each retrieval statement, for example: domain, generation time, type of knowledge content production, etc., so that each search statement can be filtered according to the filtering information after the filtering information is given.

S403, determining the requirement evaluation information of the target retrieval statement from the requirement evaluation database;

after the target search statement is determined, the requirement evaluation information of the target search statement may be further determined from a requirement evaluation database.

S404, outputting the target retrieval statement and the requirement evaluation information of the target retrieval statement.

After determining the target search term and the requirement evaluation information of the target search term, the target search term and the requirement evaluation information of the target search term may be output. Optionally, the target retrieval statement and the requirement evaluation information of the target retrieval statement may be visually displayed.

It can be understood that, besides outputting the target search statement and the requirement evaluation information of the target search statement, on the premise that other related information is recorded in the requirement evaluation database, the related information of the target search statement may also be output, for example: the ratio of the click quantity to the display quantity in the whole network, namely the result point-to-area ratio; or, the ratio of the click quantity to the display quantity under a certain domain name, namely the knowledge result point-to-area ratio; or the level of the requirement classification to which it belongs, the cluster header to which it belongs, etc.

According to the scheme provided by the disclosure, the requirement evaluation information capable of accurately reflecting the actual requirement can be generated, a realization basis is provided for generating the knowledge content matched with the actual requirement, and further, the target retrieval statement and the requirement evaluation information of the target retrieval statement can be displayed, so that a knowledge content producer can conveniently acquire required information.

According to an embodiment of the present disclosure, as shown in fig. 5, the present disclosure also provides a demand evaluation information generation apparatus, including:

a first sentence determining module 501, configured to determine a target retrieval sentence to be analyzed;

a first information obtaining module 502, configured to obtain retrieval amount information of a target retrieval statement in a first time period; the first time period is a time period before the current time, and the retrieval amount information of any retrieval statement in the time period comprises: the retrieval amount of the retrieval statement in each statistical cycle in the time period;

a pre-estimate determining module 503, configured to determine, based on the obtained retrieval amount information, an estimate corresponding to the target retrieval statement; the pre-estimated quantity is used for representing the retrieval requirement of the target retrieval statement in a second time period, and the second time period is a time period after the current time;

the information generating module 504 is configured to generate the demand evaluation information of the target search statement based on the pre-estimate corresponding to the target search statement.

Optionally, the pre-estimate determining module includes:

the data construction submodule is used for constructing the characteristic data of the target retrieval statement in a first time period based on the acquired retrieval amount information;

the model processing submodule is used for processing the characteristic data of the target retrieval statement by utilizing a pre-trained predictor prediction model to obtain a predictor corresponding to the target retrieval statement;

Optionally, the data construction sub-module is specifically configured to calculate a mean and/or a variance of each retrieval amount in the acquired retrieval amount information; and constructing a feature vector as feature data of the target retrieval statement in the first time period based on the retrieval amount included by the acquired retrieval amount information and the calculated mean value and/or variance.

Optionally, the apparatus further comprises:

the second information acquisition module is used for acquiring first retrieval amount information of the specified retrieval statement in a sample time period and second retrieval amount information of the specified retrieval statement in a labeling time period from the training data set;

the data generation module is used for generating characteristic data of the specified retrieval statement in a sample time period as sample data based on the first retrieval amount information and generating marking data of the specified retrieval statement in a marking time period based on the second retrieval amount information;

the data input module is used for inputting the sample data into the neural network model to be trained so that the neural network model can predict the prediction quantity of the specified retrieval statement in the labeling time period based on the sample data and take the prediction quantity as prediction data;

calculating a loss function value of the neural network model based on the prediction data and the labeling data;

and parameter adjustment training, namely adjusting parameters of the neural network model according to the loss function values, and performing next training until the retrieval amount information in the training sample set is utilized.

Optionally, the apparatus further comprises:

the second statement determination module is used for determining each retrieval statement existing in the sample time period;

the third information acquisition module is used for acquiring first retrieval amount information of each retrieval statement in a sample time period and second retrieval amount information in a labeling time period;

and the data set construction module is used for constructing a training data set by utilizing each retrieval statement and the first retrieval amount information and the second retrieval amount information of each retrieval statement.

Optionally, the first sentence determining module is specifically configured to filter, from among the search sentences in the first time period, the search sentences whose search frequency is greater than the preset frequency threshold as the target search sentences to be analyzed.

Optionally, the apparatus further comprises:

and the information writing module is used for writing the target retrieval statement and the requirement evaluation information of the target retrieval statement into a requirement evaluation database after the information generating module executes the step of generating the requirement evaluation information of the target retrieval statement based on the pre-estimated quantity corresponding to the target retrieval statement.

In the above scheme provided by the embodiment of the present disclosure, the estimated quantity of the target search statement in the second time period after the current time may be determined by using the search quantity information of the target search statement in the first time period before the current time, and then the demand evaluation information of the target search statement may be generated based on the determined estimated quantity. Since the prediction quantity characterizes the retrieval demand of the target retrieval statement in the second time period, the demand evaluation information can accurately reflect the actual demand of the target retrieval statement in the future second time period. Therefore, the scheme provided by the disclosure can solve the problem that the actual requirements cannot be accurately reflected by the requirement evaluation information in the related technology.

According to an embodiment of the present disclosure, as shown in fig. 6, the present disclosure further provides a demand evaluation information display apparatus, including:

a filtering information obtaining module 601, configured to obtain filtering information for a search statement;

a statement screening module 602, configured to screen a target retrieval statement matching the screening information from the retrieval statements stored in the requirement evaluation database; the demand evaluation database stores a plurality of retrieval statements and demand evaluation information of each retrieval statement, and the demand evaluation information of each retrieval statement is determined by any demand evaluation information generation device provided by the disclosure;

an information determining module 603, configured to determine, from a requirement evaluation database, requirement evaluation information of a target retrieval statement;

the information output module 604 is configured to output the target search statement and the requirement evaluation information of the target search statement.

The present disclosure also provides an electronic device, a readable storage medium, and a computer program product according to embodiments of the present disclosure.

An embodiment of the present disclosure provides an electronic device, including:

at least one processor; and

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method of implementing the demand evaluation information generation method or the demand evaluation information presentation method.

The disclosed embodiments provide a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute a method for implementing demand evaluation information generation or a demand evaluation information presentation method.

The computer program product includes a computer program, and the computer program, when executed by a processor, implements a demand evaluation information generating method or a demand evaluation information displaying method.

FIG. 7 illustrates a schematic block diagram of an example electronic device 700 that can be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers

. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 7, the device 700 comprises a computing unit 701, which may perform various suitable actions and processes according to a computer program stored in a Read Only Memory (ROM)702 or a computer program loaded from a storage unit 708 into a Random Access Memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the device 700 can also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other by a bus 704. An input/output (I/O) interface 705 is also connected to bus 704.

Various components in the device 700 are connected to the I/O interface 705, including: an input unit 706 such as a keyboard, a mouse, or the like; an output unit 707 such as various types of displays, speakers, and the like; a storage unit 708 such as a magnetic disk, optical disk, or the like; and a communication unit 709 such as a network card, modem, wireless communication transceiver, etc. The communication unit 709 allows the device 700 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.

Computing unit 701 may be a variety of general purpose and/or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and so forth. The calculation unit 701 performs each method and process described above, for example, implements a demand evaluation information generation method or a demand evaluation information presentation method. For example, in some embodiments, implementing the demand evaluation information generation method or the demand evaluation information presentation method may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as storage unit 708. In some embodiments, part or all of a computer program may be loaded onto and/or installed onto device 700 via ROM 702 and/or communications unit 709. When loaded into the RAM 703 and executed by the computing unit 701, may perform one or more of the steps described above to implement the requirement evaluation information generation method or the requirement evaluation information presentation method. Alternatively, in other embodiments, the computing unit 701 may be configured by any other suitable means (e.g., by means of firmware) to perform implementing the requirement evaluation information generation method or the requirement evaluation information presentation method.

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowchart and/or block diagram to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.

The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server with a combined blockchain.

It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure may be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and the present disclosure is not limited herein.

The above detailed description should not be construed as limiting the scope of the disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. A demand evaluation information generation method includes:

determining a target retrieval statement to be analyzed;

2. The method according to claim 1, wherein the determining the pre-estimate amount corresponding to the target search statement based on the acquired search amount information comprises:

constructing feature data of the target retrieval statement in the first time period based on the acquired retrieval amount information;

processing the characteristic data of the target retrieval statement by using a pre-trained predictor prediction model to obtain a predictor corresponding to the target retrieval statement;

wherein, the prediction model is as follows: the method comprises the steps that training is carried out on the basis of sample data and labeled data corresponding to the sample data, wherein the sample data is characteristic data of a specified retrieval statement in a sample time period, the labeled data is used for representing retrieval requirements of the specified retrieval statement in a labeled time period, the sample time period is a time period before reference time, and the labeled time period is a time period after the reference time.

3. The method according to claim 2, wherein the constructing feature data of the target retrieval statement in the first time period based on the acquired retrieval amount information includes:

calculating the mean value and/or the variance of each retrieval amount in the acquired retrieval amount information;

and constructing a feature vector as feature data of the target retrieval statement in the first time period based on the retrieval amount included by the acquired retrieval amount information and the calculated mean value and/or variance.

4. The method of claim 2, wherein the predictor predictive model is trained in the following manner:

acquiring first retrieval amount information of a specified retrieval statement in the sample time period and second retrieval amount information of the specified retrieval statement in the labeling time period from a training data set;

generating feature data of the specified retrieval statement in the sample time period as sample data based on the first retrieval amount information, and generating annotation data of the specified retrieval statement in the annotation time period based on the second retrieval amount information;

inputting the sample data into a neural network model to be trained, so that the neural network model predicts the prediction quantity of the specified retrieval statement in the labeling time period based on the sample data, and the prediction quantity is used as prediction data;

calculating a loss function value of the neural network model based on the prediction data and the annotation data;

and adjusting the neural network model parameters according to the loss function values, and carrying out next training until the retrieval amount information in the training sample set is utilized.

5. The method of claim 4, wherein the training data set is constructed in a manner comprising:

determining each retrieval statement existing in the sample time period;

acquiring first retrieval amount information of each retrieval statement in the sample time period and second retrieval amount information in the labeling time period;

and constructing a training data set by using each retrieval statement and the first retrieval amount information and the second retrieval amount information of each retrieval statement.

6. The method of any of claims 1-5, wherein the determining a target retrieval statement to be analyzed comprises:

and screening the retrieval sentences of which the retrieval frequency is greater than a preset frequency threshold value from the retrieval sentences in the first time period to serve as target retrieval sentences to be analyzed.

7. The method according to any one of claims 1 to 5, further comprising, after the generating demand evaluation information of the target search statement based on the corresponding pre-estimated amount of the target search statement,:

8. A method for displaying demand evaluation information, the method comprising:

acquiring screening information aiming at the retrieval statement;

screening target retrieval sentences matched with the screening information from the retrieval sentences stored in the requirement evaluation database; wherein, a plurality of retrieval statements and requirement evaluation information of each retrieval statement are stored in the requirement evaluation database, and the requirement evaluation information of each retrieval statement is determined according to the method of any one of claims 1 to 7;

determining requirement evaluation information of the target retrieval statement from the requirement evaluation database;

and outputting the target retrieval statement and the requirement evaluation information of the target retrieval statement.

9. A demand evaluation information generation apparatus comprising:

10. The apparatus of claim 9, wherein the prediction determination module comprises:

the data construction sub-module is used for constructing the characteristic data of the target retrieval statement in the first time period based on the acquired retrieval amount information;

11. The apparatus according to claim 10, wherein the data construction sub-module is specifically configured to calculate a mean and/or a variance of each retrieval amount in the acquired retrieval amount information; and constructing a feature vector as feature data of the target retrieval statement in the first time period based on the retrieval amount included by the acquired retrieval amount information and the calculated mean value and/or variance.

12. The apparatus of claim 10, the apparatus further comprising:

the second information acquisition module is used for acquiring first retrieval amount information of a specified retrieval statement in the sample time period and second retrieval amount information of the specified retrieval statement in the labeling time period from a training data set;

a data generating module, configured to generate, based on the first retrieval amount information, feature data of the specified retrieval statement in the sample time period as sample data, and generate, based on the second retrieval amount information, labeled data of the specified retrieval statement in the labeled time period;

the data input module is used for inputting the sample data into a neural network model to be trained so that the neural network model can predict the prediction quantity of the specified retrieval statement in the labeling time period based on the sample data to be used as prediction data;

and parameter adjustment training, namely adjusting the parameters of the neural network model according to the loss function values, and performing next training until the retrieval amount information in the training sample set is utilized.

13. The apparatus of claim 12, the apparatus further comprising:

the third information acquisition module is used for acquiring first retrieval amount information of each retrieval statement in the sample time period and second retrieval amount information in the labeling time period;

14. The apparatus according to any one of claims 9 to 13, wherein the first sentence determination module is specifically configured to filter, from among the search sentences in the first time period, search sentences whose search frequency is greater than a preset frequency threshold as target search sentences to be analyzed.

15. The apparatus of any of claims 9-13, wherein the apparatus further comprises:

and the information writing module is used for writing the target retrieval statement and the requirement evaluation information of the target retrieval statement into a requirement evaluation database after the information generating module executes the step of generating the requirement evaluation information of the target retrieval statement based on the pre-estimation amount corresponding to the target retrieval statement.

16. A demand evaluation information presentation apparatus, the apparatus comprising:

the statement screening module is used for screening target retrieval statements matched with the screening information from the retrieval statements stored in the requirement evaluation database; wherein the requirement evaluation database stores a plurality of search sentences and requirement evaluation information of each search sentence, the requirement evaluation information of each search sentence being determined by the apparatus according to any one of claims 9 to 15;

17. An electronic device, comprising:

at least one processor; and

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7 or 8.

18. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any of claims 1-7 or 8.

19. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-7 or 8.