CN114661890A

CN114661890A - Knowledge recommendation method, device, system and storage medium

Info

Publication number: CN114661890A
Application number: CN202210344905.5A
Authority: CN
Inventors: 徐凯旋; 张祖亮
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2022-03-31
Filing date: 2022-03-31
Publication date: 2022-06-24

Abstract

The present disclosure provides a knowledge recommendation method, apparatus, system and storage medium, which relate to the technical field of artificial intelligence, and in particular to the technical field of big data, Natural Language Processing (NLP) and intelligent customer service. The specific implementation scheme is as follows: acquiring a problem to be solved submitted by a terminal device; acquiring first knowledge for solving a problem to be solved from a preset knowledge base; determining an association score of each other knowledge and the first knowledge based on the theme of the first knowledge and the themes of the other knowledge in the preset knowledge base; determining second knowledge associated with the first knowledge from the other knowledge according to the association score; and recommending the title of the second knowledge and the first knowledge to the terminal equipment, wherein the first knowledge is an answer of the problem to be solved, and the title of the second knowledge is a related problem of the problem to be solved.

Description

Knowledge recommendation method, device, system and storage medium

Technical Field

The present disclosure relates to the technical field of artificial intelligence, particularly to the technical field of big data, Natural Language Processing (NLP) and intelligent customer service, and more particularly to a method, an apparatus, a system and a storage medium for knowledge recommendation.

Background

With the development and landing of AI (Artificial Intelligence) technology, intelligent customer service becomes a standard solution to replace Artificial customer service. The knowledge recommendation system is an important component of intelligent customer service and is used for recommending a solution to a problem for a user.

Disclosure of Invention

The disclosure provides a knowledge recommendation method, a knowledge recommendation device, a knowledge recommendation system and a storage medium.

According to an aspect of the present disclosure, there is provided a knowledge recommendation method including:

acquiring a problem to be solved submitted by a terminal device;

acquiring first knowledge for solving the problem to be solved from a preset knowledge base;

determining an association score of each other knowledge with the first knowledge based on the topic of the first knowledge and the topics of the other knowledge in the preset knowledge base;

determining second knowledge associated with the first knowledge from the other knowledge based on the association score;

and recommending the title of the second knowledge and the first knowledge to the terminal equipment, wherein the first knowledge is an answer of the problem to be solved, and the title of the second knowledge is a related problem of the problem to be solved.

According to another aspect of the present disclosure, there is provided a knowledge recommendation apparatus including:

the first acquisition unit is used for acquiring the problem to be solved submitted by the terminal equipment;

a second acquiring unit, configured to acquire, from a preset knowledge base, first knowledge for solving the problem to be solved;

the first determination unit is used for determining the association score of each other knowledge and the first knowledge based on the theme of the first knowledge and the themes of other knowledge in the preset knowledge base;

a second determining unit, configured to determine, according to the association score, second knowledge associated with the first knowledge from the other knowledge;

and the recommending unit is used for recommending the title of the second knowledge and the first knowledge to the terminal equipment, wherein the first knowledge is an answer of the problem to be solved, and the title of the second knowledge is a related problem of the problem to be solved.

According to a third aspect of the present disclosure, there is provided a knowledge recommendation system comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein, the first and the second end of the pipe are connected with each other,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform any of the above-described knowledge recommendation methods.

According to a fourth aspect of the present disclosure, there is provided a non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform any of the above knowledge recommendation methods.

According to a fifth aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements a method of recommendation according to any of the above knowledge.

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 first flowchart of a knowledge recommendation method provided by an embodiment of the disclosure;

FIG. 2 is a diagram of a dialog box for associating knowledge provided by an embodiment of the present disclosure;

FIG. 3 is a second flowchart of a knowledge recommendation method provided by an embodiment of the disclosure;

FIG. 4 is a diagram of a dialog box for hot question recommendation provided by an embodiment of the present disclosure;

FIG. 5 is a third flowchart of a knowledge recommendation method provided by an embodiment of the disclosure;

FIG. 6 is a diagram of a dialog box for candidate headline recommendation provided by an embodiment of the present disclosure;

FIG. 7 is a fourth flowchart illustrating a knowledge recommendation method according to an embodiment of the disclosure;

FIG. 8 is a fifth flowchart of a knowledge recommendation method provided by an embodiment of the disclosure;

FIG. 9 is a sixth flowchart of a knowledge recommendation method provided by an embodiment of the disclosure;

FIG. 10 is a seventh flowchart of a knowledge recommendation method provided by an embodiment of the disclosure;

FIG. 11 is a diagram of a dialog box for a directory question provided by an embodiment of the present disclosure;

FIG. 12 is an eighth flowchart illustrating a knowledge recommendation method according to an embodiment of the disclosure;

FIG. 13 is a ninth flowchart illustrating a knowledge recommendation method according to an embodiment of the disclosure;

FIG. 14 is a tenth flowchart illustrating a knowledge recommendation method according to an embodiment of the disclosure;

FIG. 15 is an eleventh flowchart illustrating a knowledge recommendation method according to an embodiment of the disclosure;

FIG. 16 is a schematic diagram of a metamorphic image provided by embodiments of the present disclosure;

FIG. 17 is a twelfth flowchart illustrating a knowledge recommendation method according to an embodiment of the disclosure;

FIG. 18 is a schematic diagram of a knowledge recommendation device provided by an embodiment of the disclosure;

FIG. 19 is a first block diagram of a knowledge recommendation system for implementing the knowledge recommendation method of an embodiment of the present disclosure;

FIG. 20 is a second block diagram of a knowledge recommendation system for implementing the knowledge recommendation method of 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.

The knowledge recommendation system is an important component of intelligent customer service and is used for recommending a solution to a problem for a user. At present, a knowledge recommendation system recommends a solution to a problem to a user based on manual configuration, specifically: the knowledge recommendation system receives a problem to be solved input by a user, determines target knowledge corresponding to the problem to be solved from a pre-stored corresponding relation between the problem and the knowledge, acquires pre-configured associated knowledge of the target knowledge, forms a related problem list by the title of the associated knowledge, and recommends the target knowledge and the related problem list to the user.

The above knowledge recommendation method has the following problems:

first, the adaptation scenario is single and only supports knowledge recommendation in the knowledge correlation problem scenario.

Second, it is necessary to manually associate the relevant knowledge for each piece of knowledge, and the larger the amount of knowledge, the more manpower consumed.

Thirdly, the relevant knowledge recommended by manual association has one-sidedness, and particularly for a large knowledge base with the knowledge level of more than 10 thousands, all knowledge in the knowledge base is difficult to master manually, the knowledge relevant to each knowledge is difficult to be associated globally, and further the relevant knowledge is difficult to be recommended to the user globally.

To solve the above problem, the embodiment of the present disclosure provides a knowledge recommendation method, which may be applied to a knowledge recommendation system, as shown in fig. 1. The knowledge recommendation system can be located on a terminal device for inputting problems by a user, wherein the terminal device is a device for performing human-Computer interaction with the user and can be an electronic device such as a mobile phone, a Personal Computer (PC), a tablet Computer and the like. The knowledge recommendation system can also be an independent physical machine, such as a knowledge recommendation system and the like. For ease of understanding, the following description will use the knowledge recommendation system as the execution subject, and is not intended to be limiting. The knowledge recommendation method comprises the following steps:

and step S11, acquiring the problem to be solved submitted by the terminal equipment.

In the embodiment of the disclosure, the user can input the problem to be solved, i.e. the problem to be solved, to the terminal device. The terminal equipment submits the problems to be solved input by the user to the knowledge recommendation system, and then the knowledge recommendation system obtains the problems to be solved submitted by the terminal equipment.

In step S12, first knowledge for solving the problem to be solved is acquired from a preset knowledge base.

In the embodiment of the present disclosure, the preset knowledge base includes a plurality of knowledge, and the preset knowledge base may be implemented by using a MySQL (structured query language) database, an ES (distributed full text search engine) database, or a database in other forms, which is not limited herein.

After the knowledge recommendation system obtains the problem to be solved, the knowledge recommendation system searches in a preset knowledge base to obtain an answer, namely first knowledge, of the problem to be solved.

Step S13, determining the association score of each other knowledge and the first knowledge based on the theme of the first knowledge and the themes of other knowledge in the preset knowledge base.

In the disclosed embodiment, each knowledge in the predetermined knowledge base may include one or more topics.

The knowledge recommendation system extracts the theme of each knowledge in the preset knowledge base, namely extracts the theme of the first knowledge, and extracts other knowledge themes in the preset knowledge base. For each other knowledge, the knowledge recommendation system determines an association score for the other knowledge with the first knowledge based on the topic of the first knowledge and the topic of the other knowledge. A larger association score indicates a higher degree of association of the other knowledge with the first knowledge, and a smaller association score indicates a lower degree of association of the other knowledge with the first knowledge.

In step S14, the second knowledge related to the first knowledge is determined from the other knowledge according to the relevance score.

After obtaining the association score of each of the other knowledge and the first knowledge, the knowledge recommendation system may determine, as the second knowledge, the other knowledge associated with the first knowledge according to the magnitude of each of the association scores.

In an alternative embodiment, the knowledge recommendation system may determine a sixth preset number of other knowledge with the highest association score from the other knowledge, and further use the sixth preset number of other knowledge as the second knowledge associated with the first knowledge.

The sixth preset number may be set according to actual requirements, for example, the sixth preset number may be 5, 8, or 10, and is not limited thereto.

In another alternative embodiment, the knowledge recommendation system may determine other knowledge having an association score higher than a preset score threshold from the other knowledge, and use the other knowledge as the second knowledge associated with the first knowledge. The preset scoring threshold value can be set according to actual requirements, and is not limited.

In the technical solution provided by the embodiment of the present disclosure, the association score of the other knowledge and the first knowledge is proportional to the association degree, that is, the knowledge recommendation system may obtain a limited number of the knowledge with the highest association degree with the first knowledge, that is, the second knowledge. If the knowledge recommendation system recommends the second knowledge to the terminal equipment, the user can better solve the problem to be solved according to the second knowledge with the highest association degree through the terminal equipment, and the problem solving efficiency of the user is improved. Meanwhile, only limited second knowledge is recommended to the user, so that resources of the knowledge recommendation method can be saved, and invalid resource consumption is avoided.

The knowledge recommendation system may also determine the second knowledge by other methods, which are not limited.

And step S15, recommending the title of the second knowledge and the first knowledge to the terminal equipment, wherein the first knowledge is an answer of the problem to be solved, and the title of the second knowledge is a related problem of the problem to be solved.

In the embodiment of the present disclosure, each knowledge in the preset knowledge base has a corresponding title, and the title may be represented as a question corresponding to the knowledge.

After the first knowledge and the second knowledge are obtained, the knowledge recommendation system takes the first knowledge as an answer of the problem to be solved and takes the title of the second knowledge as a related problem of the problem to be solved. And then the knowledge recommendation system recommends answers to the problems to be solved and related problems to the terminal equipment. And after receiving the answer and the related problems of the problem to be solved, the terminal equipment displays the answer and the related problems of the problem to be solved.

As shown in the dialog box related to knowledge in fig. 2, the user X inputs the question 1 in the input interface of the dialog box and submits the question 1 to the knowledge recommendation system, as shown in the question 1 of the user X on the right side of the display interface in fig. 2, and subsequently, after receiving the answer to the question 1 and the related question of the question 1, the terminal device displays the answer to the question 1 and the related question of the question 1 recommended by the knowledge recommendation system, as shown in the information displayed under the intelligent assistant on the left side of the display interface in fig. 2.

In the embodiment of the disclosure, the knowledge recommendation system may rank the plurality of second knowledge according to the order of the relevance scores from large to small, so that the knowledge recommendation system may form a recommendation list of related problems according to the titles of the ranked second knowledge, and recommend the recommendation list of related problems to the terminal device, so that the user can obtain the knowledge of interest from the recommendation list, and the efficiency of the user in solving the problems is improved.

In the technical solution provided by the embodiment of the present disclosure, after acquiring the first knowledge, the knowledge recommendation system may determine, based on the topic of each other knowledge and the topic of the first knowledge in the preset knowledge base, an association score of each other knowledge and the first knowledge. Compared with the traditional method for determining the association score based on the keywords, the knowledge recommendation method for determining the association score based on the topics has higher accuracy, and particularly under the condition that other words with similar meaning and different shapes exist in the knowledge and the first knowledge, the association relationship among the words cannot be judged by the method for determining the association score based on the keywords, and the knowledge recommendation method for determining the association score based on the topics can avoid the condition.

In addition, the knowledge recommendation system may determine second knowledge associated with the first knowledge based on the association score after determining the association score, and recommend a title of the second knowledge and the first knowledge to the terminal device, that is, to a user using the terminal device. In this way, the user can solve the problem to be solved based on the first knowledge. If the first knowledge cannot solve the problem to be solved, the user can also check the title of the recommended second knowledge, namely, the related problems of the problem to be solved are checked, the problem to be solved is solved by checking the related problems, and the problem solving efficiency of the user is improved.

In addition, the knowledge recommendation system can determine the second knowledge associated with the first knowledge according to the association scores to further determine the related problems, the related knowledge does not need to be configured manually, manpower consumption is reduced, and the knowledge recommendation system can accurately compare the first knowledge with all knowledge in the preset knowledge base to determine the second knowledge associated with the first knowledge even if the knowledge magnitude stored in the knowledge base is large, so that the related problems are recommended accurately, and the comprehensiveness of the related problems is improved.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, and as shown in fig. 3, the method may include steps S31-S37, and steps S34-S37 are the same as steps S12-S15, and are not described herein again. Steps S31-S33 are one possible implementation of step S11.

And step S31, acquiring the question of the question type from the question pool of the question type according to the preset weight of each question type, and taking the question as the hot question.

In the embodiment of the present disclosure, the problem type may include at least one of the following: a type of first high frequency question, a type of second high frequency question and a type of preset question.

Wherein, the first high-frequency problem is: and the titles of the knowledge with the highest access times in the preset time length and the first preset number are accessed. The preset time period may be set according to actual requirements, for example, 12 hours, 1 day, 3 days, or 5 days, which is not limited; the first preset number may also be set according to actual requirements, which is not limited. The number of access times of the knowledge is the total number of times that the knowledge recommendation system recommends the knowledge to each terminal device, namely, the number of access times of the knowledge is increased by 1 when the knowledge recommendation system recommends one knowledge to the terminal device at a time.

The second high frequency problem is: and in the knowledge associated with the knowledge recommended to the terminal equipment, titles of a second preset number of knowledge with the highest access times within a preset time length. The second preset quantity can be set according to actual requirements, and the first preset quantity and the second preset quantity can be the same or different and are not limited. Here, the determination of the associated knowledge may be as described above in steps S13-S14 with respect to determining the second knowledge associated with the first knowledge.

Preset questions may include, but are not limited to, questions that the administrator configures in the background based on new products released, new functions, and promotional activities.

In the embodiment of the present disclosure, the problem types may also include other types, and the problems of different problem types may be the same or different, which is not limited herein.

The knowledge recommendation system may preset a corresponding weight for each question type, i.e., a preset weight. The preset weight of each problem type can be set according to actual requirements, for example: the preset weight of the question type 1 is 3, the preset weight of the question type 2 is 3, and the preset weight of the question type 3 is 4, which is not limited.

In the embodiment of the disclosure, the knowledge recommendation system may randomly acquire the problem of the problem type from the problem pool of the problem type according to the preset weight of each problem type, as a hot problem.

For example, the knowledge recommendation system sets a problem pool of problem types 1-3, where the preset weight of the problem type 1 is 3, the preset weight of the problem type 2 is 3, and the preset weight of the problem type 3 is 4, and 10 hot problems are required to be obtained in total, and then the knowledge recommendation system obtains 3 problems of the problem type 1 from the problem pool of the problem type 1 as the hot problems; acquiring 3 questions of the question type 2 from the question pool of the question type 2 as hot questions; from the problem pool of problem type 3, 4 problems of problem type 3 are obtained as the hot problem.

In step S32, the obtained plurality of hot questions are recommended to the terminal device, so that the terminal device determines a target hot question from the plurality of hot questions.

The knowledge recommendation system can recommend the acquired hot problems to the terminal equipment. After receiving the plurality of hot questions, the terminal device may randomly select one of the plurality of hot questions as a target hot question.

The terminal device may also output the plurality of hot questions after receiving the plurality of hot questions. Fig. 4 shows 10 hit questions, wherein the number of hit questions in the figure is merely an example and is not limited to the dialog box for hit question recommendation shown in fig. 4.

The user can select one hot problem from the hot problems displayed by the terminal equipment; and after detecting the operation of selecting the hot problem by the user, the terminal equipment determines that the hot problem selected by the user is the target hot problem.

And step S33, receiving the target hot problem determined by the terminal device as the problem to be solved.

After the knowledge recommendation system receives the target hot problem determined by the terminal device, the knowledge recommendation system can take the target hot problem as a problem to be solved submitted by the terminal device.

In the technical scheme provided by the embodiment of the disclosure, the knowledge recommendation system can acquire the problem of each problem type according to the preset weight of the problem type, and the problem is taken as a hot problem. And recommending the obtained hot problems to the terminal equipment to determine the problems to be solved. Therefore, the knowledge recommendation system can recommend knowledge to the user according to different problem types, the idea of solving the problems of the user is expanded from different angles, and the efficiency of solving the problems of the user is improved.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, and as shown in fig. 5, the method may include steps S51-S59, and steps S56-S59 are the same as steps S12-S15, and are not described herein again. Steps S51-S55 are one possible implementation of step S11.

And step S51, acquiring candidate texts submitted by the terminal equipment.

In the embodiment of the present disclosure, the candidate text may be one or more keywords, or may be a complete problem to be solved, which is not limited herein. The user can submit the candidate text to the knowledge recommendation system through the terminal equipment, and the knowledge recommendation system obtains the candidate text submitted by the terminal equipment.

In step S52, a plurality of candidate titles matching the candidate text are obtained from the titles of knowledge included in the preset knowledge base.

The knowledge recommendation system can perform matching retrieval on the titles of the knowledge included in the preset knowledge base and the candidate texts submitted by the terminal equipment, and acquire a plurality of candidate titles matched with the candidate texts from the titles of the knowledge included in the preset knowledge base.

In step S53, a similarity score between each candidate title and the candidate text is calculated.

For each candidate title, the knowledge recommendation system may calculate a similarity score between the candidate title and the candidate text by adopting a TF-IDF (Term-Inverse Document Frequency) algorithm or other algorithms for calculating text similarity.

In step S54, recommending a third preset number of candidate titles with the highest similarity scores to the terminal device, so that the terminal device determines the target association problem from the third preset number of candidate titles.

The knowledge recommendation system may select a third preset number of candidate titles with the highest similarity scores from the obtained plurality of candidate titles, and recommend the candidate titles to the terminal device. The third preset number may be set according to actual requirements, for example, the third preset number may be 5, 8, or 10, which is not limited.

After receiving the plurality of candidate titles, the terminal device may randomly determine one candidate title as the target association problem from a third preset number of candidate titles.

The terminal device may also output a plurality of candidate titles after receiving the plurality of candidate titles. As shown in fig. 6, in the dialog box for recommending candidate titles, in fig. 6, after the input interface inputs candidate texts, 5 candidate titles are presented above the input interface, where the number of candidate titles in the figure is merely an example and is not limited.

The user can select one candidate title from the candidate titles displayed by the terminal equipment; after detecting the operation of selecting the candidate title by the user, the terminal equipment determines that the candidate title selected by the user is the target associated title.

In the embodiment of the disclosure, the knowledge recommendation system ranks the third preset number of candidate titles in the order from high to low of the similarity score, and recommends the third preset number of candidate titles to the terminal device according to the ranking result. The higher the similarity score is, the higher the probability that the candidate title is interested by the user is, and the candidate title is recommended to the terminal equipment according to the sequence of the similarity score from high to low, so that the user can conveniently obtain the interested knowledge from the recommendation list, and the problem solving efficiency of the user is improved.

In step S55, the target association problem determined by the terminal device is received as the problem to be solved.

After the knowledge recommendation system receives the target association problem determined by the terminal device, the knowledge recommendation system can take the target association problem as a problem to be solved submitted by the terminal device.

In the technical scheme provided by the embodiment of the disclosure, the knowledge recommendation system can acquire a plurality of candidate titles matched with the candidate texts according to the candidate texts submitted by the terminal equipment, and recommend the acquired third preset number of candidate titles to the terminal equipment so as to determine the problem to be solved. Therefore, the user can select the candidate title matched with the candidate text under the condition that the complete problem is not submitted, the problem to be solved is solved by checking the knowledge corresponding to the selected candidate title, and the problem solving efficiency of the user is improved.

And the knowledge recommendation system matches the candidate titles once every time the user inputs a keyword, and the candidate titles matched with the knowledge recommendation system are changed along with the increase and change of the keyword in the candidate text input by the user. Therefore, under the condition that the user is unclear about the problem to be solved, the knowledge recommendation system can support the user to input a single keyword or change the search of the keyword at any time so as to expand the idea of solving the problem by the user and further improve the efficiency of solving the problem by the user; when the user has a definite question asking direction for the problem to be solved, the knowledge recommendation system can also support the user to input a plurality of keywords or input a complete problem, so that the knowledge recommendation system can feed back the answer for solving the problem to the user more accurately.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, and as shown in fig. 7, the method may include steps S71-S710, where steps S71 and steps S74-S710 are the same as steps S51 and steps S53-S59, and are not described herein again. Steps S72-S73 are one possible implementation of step S52.

And step S72, performing word segmentation on the candidate text to obtain a plurality of candidate word segmentations.

In the embodiment of the disclosure, the knowledge recommendation system can perform word segmentation on the candidate text to obtain a plurality of candidate word segmentations.

If the preset knowledge base is stored in the ES, the knowledge recommendation system may perform word segmentation on the candidate text by using a chinese word segmenter such as ik _ smart (coarsest force split) or ik _ max _ word (finest force split), so as to obtain a plurality of candidate word segmentations. The knowledge recommendation system may also adopt other methods for segmenting the candidate text, which is not limited herein.

In step S73, a title including the candidate participle is acquired as a candidate title from titles of knowledge included in the preset knowledge base.

In the embodiment of the disclosure, the knowledge recommendation system may query the titles of the knowledge included in the preset knowledge base, acquire the titles including the candidate participles, and use the titles as the candidate titles.

In the embodiment of the present disclosure, the number of candidate participles may be one or more, and if the number of candidate participles is multiple and one candidate participle is included in one title, the candidate participle may be used as a candidate title.

In the technical scheme provided by the embodiment of the disclosure, the knowledge recommendation system can perform word segmentation on the candidate text and acquire a title comprising the candidate word segmentation as the candidate title. Because the candidate text may be a complete problem to be solved, the candidate text is subjected to word segmentation, so that the flexibility of the obtained candidate titles can be improved, and the flexibility of knowledge recommendation is further improved.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, and as shown in fig. 8, the method may include steps S81-S811, and steps S81-S83 and steps S86-S811 are the same as steps S71-S73 and steps S75-S710, and are not described herein again. Steps S84-S85 are one possible implementation of step S74.

Step S84, calculating a ratio of a first coefficient of each candidate participle associated with each candidate heading to a second coefficient of the candidate participle, where the first coefficient of each candidate participle is the number of occurrences of the candidate participle in the candidate heading, and the second coefficient of each candidate participle is the number of candidate headings including the candidate participle.

In the embodiment of the present disclosure, after obtaining the candidate titles, for each candidate participle, the knowledge recommendation system may determine the occurrence number of the candidate participle in the candidate title, as a first coefficient of the candidate participle associated with the candidate title, determine the number of candidate titles including the candidate participle, as a second coefficient of the candidate participle associated with the candidate title; and calculating the ratio of the first coefficient of the candidate participle related to the candidate title to the second coefficient of the candidate participle.

This ratio can also be considered as: and calculating the similarity score of the candidate title and the candidate text by adopting a TF-IDF method to obtain a TF-IDF value.

In step S85, for each candidate heading, the sum of the ratios of all candidate participles associated with the candidate heading is used as the similarity score between the candidate heading and the candidate text.

In the embodiment of the disclosure, for each candidate title, the knowledge recommendation system may use the sum of the ratios (i.e., TF-IDF values) of all candidate participles associated with the candidate title as the similarity score of the candidate title and the candidate text.

In the technical scheme provided by the embodiment of the disclosure, for each candidate title, the knowledge recommendation system may obtain a sum of ratios of all candidate participles associated with the candidate title as a similarity score of the candidate title and the candidate text. Therefore, the knowledge recommendation system can determine the association relation between each candidate title and each candidate participle, so as to determine the similarity score between the candidate title and the candidate text, improve the probability that the candidate title is the problem interested by the user, and improve the efficiency of solving the problem by the user.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, as shown in fig. 9, in the method, there are multiple query algorithms for the knowledge recommendation system to obtain candidate titles, and the query algorithms may include, but are not limited to, a match query, a match _ phrase query, and the like.

Optionally, when the knowledge recommendation system queries the title of the knowledge included in the preset knowledge base, the knowledge recommendation system may further add the professional term or the product name included in the input text to the ES dictionary, and then query the title of the knowledge included in the preset knowledge base to obtain the title including the candidate word segmentation as the candidate title, so that the accuracy of obtaining the candidate title may be improved, and the accuracy of the knowledge recommendation method may be further improved.

The knowledge recommendation method shown in FIG. 9 may include steps S91-S910, and steps S91, S92 and S95-S910 are the same as the above steps S51, S52 and S54-S59, and are not described herein again. Steps S93-S94 are one possible implementation of step S53.

Step S93, for each query algorithm, calculating an initial similarity score between each candidate title and each candidate text obtained by using the query algorithm.

For each query algorithm, the knowledge recommendation system may calculate an initial similarity score corresponding to the query algorithm, which may refer to the related description of the similarity score obtained by calculation in fig. 8, and details are not repeated here.

In step S94, for each candidate title, the sum of the initial similarity scores of the candidate title is calculated as the similarity score between the candidate title and the candidate text.

For each candidate heading, the knowledge recommendation system may calculate a sum of a plurality of initial similarity scores for the candidate heading as the similarity score for the candidate heading and the candidate text.

In the technical scheme provided by the embodiment of the disclosure, the knowledge recommendation system can adopt a plurality of query algorithms to calculate the similarity score and acquire the candidate titles to be recommended to the user, so that the defects of different query algorithms can be mutually compensated, and the flexibility and the accuracy of recommending the candidate titles are improved.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, as shown in fig. 10, the method may include steps S101 to S108, and steps S105 to S108 are the same as steps S12 to S15, and are not described herein again. Steps S101-S104 are one possible implementation of step S11.

Step S101, recommending a plurality of hot catalogs to the terminal equipment so that the terminal equipment displays the hot catalogs in a navigation mode, and determining a target catalog from the hot catalogs, wherein the hot catalogs are the catalogs which comprise a plurality of knowledge and have the access times and values higher than a preset access threshold value.

The knowledge recommendation system presets an access threshold, namely a preset access threshold. The knowledge in the preset knowledge base belongs to a plurality of catalogues, and each catalogue comprises a plurality of knowledge. The knowledge recommendation system counts the access times and values of a plurality of pieces of knowledge under each directory in a preset knowledge base, further obtains the directory with the access times and values higher than a preset access threshold value as a hot directory, and recommends the hot directory to the terminal equipment.

After receiving the hit list, the terminal device displays the hit list in a navigation form, such as a dialog box shown in fig. 4, and in fig. 4, the bottom of the display interface displays a plurality of hit lists, such as lists 1-5, where the number of candidate titles in the figure is merely an example and is not limited. In fig. 4, an update button is provided on the right side of the navigated directories 1-5, and when there is no directory of interest to the user in the currently presented hot directory, the user can click the update button to update the navigated hot directory.

The user can select one hot catalog from the hot catalogs displayed by the terminal equipment; and after the terminal equipment detects the operation of selecting the hot directory by the user, determining the hot directory selected by the user as the target directory.

In order to improve the knowledge recommendation efficiency, the knowledge recommendation system can sort a plurality of popular directories according to the sequence of the access times and values corresponding to the popular directories from large to small, and recommend the sorting result to the terminal device, so that a user can obtain the interested directories from the sorting result, and the knowledge recommendation efficiency is improved.

And step S102, receiving the target directory submitted by the terminal equipment.

And after the terminal equipment determines the target directory, submitting the target directory to a knowledge recommendation system.

And step S103, recommending a fourth preset number of knowledge titles with the highest access frequency under the target directory to the terminal equipment so that the terminal equipment can determine the target directory problem from the titles under the target directory.

In the embodiment of the present disclosure, the fourth preset number may be set according to an actual requirement, for example, the fourth preset number may be 5, 8, or 10, which is not limited herein.

The knowledge recommendation system can determine a fourth preset number of knowledge with the highest access times from a plurality of knowledge in the target directory, recommend the titles of the fourth preset number of knowledge to the terminal device, and display the titles of the fourth preset number of knowledge after the terminal device receives the titles of the fourth preset number of knowledge. As shown in the dialog box of the directory question in fig. 11, after the terminal device submits the directory 1, the area 1 and the area 2 in the display interface respectively display the titles of the plurality of knowledge under the directory 1. In the embodiment of the present disclosure, the terminal device may also display the titles of the fourth preset amount of knowledge in other manners, for example, only the titles of the fourth preset amount of knowledge are displayed in the area 1 or the area 2 in fig. 11. In fig. 11, an update button is provided on the right side of the navigated directory question 7-10, and when there is no question of interest to the user in the currently presented directory question, the user can click the update button to update the navigated directory question.

In order to improve the efficiency of solving problems for users, the knowledge recommendation system can sort a plurality of popular directories according to the sequence of the access times and values corresponding to the popular directories from large to small, and recommend the sorting result to the terminal device, so that the users can obtain interested problems from the sorting result, and the efficiency of recommending knowledge is improved.

Optionally, if the terminal device includes multiple areas for displaying the problems, the terminal device may configure the priority of each area, the area with the higher priority, and the larger the access times and value of the hot directory corresponding to the display problem are, the more the problems displayed in each area are sorted from the largest to the smallest access times of the knowledge. In this way, the efficiency of the user in solving the problem can be further improved. For example, as shown in FIG. 11, area 1 and area 2 are both used to present questions, where area 1 has a higher priority than area 2, i.e., directory questions 1-6 presented in area 1 have a higher number of accesses than directory questions 7-9 presented in area 2.

And step S104, receiving the target directory problem determined by the terminal equipment as the problem to be solved.

The knowledge recommendation system can receive the target directory problem determined by the terminal device and take the target directory problem as a problem to be solved.

In the technical scheme provided by the embodiment of the disclosure, the knowledge recommendation system can recommend problems for the user based on the hot directory and the knowledge with the highest access times under the hot directory, so that the efficiency of determining the problems to be solved is improved, and the problem solving efficiency of the user is further improved.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, as shown in fig. 12, the method may include steps S121 to S128, and steps S121 and S126 to S128 are the same as steps S11 and steps S13 to S15, and are not described herein again. Steps S122-S125 are one implementation of step S12.

And S122, performing word segmentation on the problem to be solved to obtain a plurality of target word segments.

In the embodiment of the disclosure, the knowledge recommendation system can perform word segmentation on the problem to be solved to obtain a plurality of target word segments. The knowledge recommendation system can adopt Chinese word segmenters such as ik _ smart or ik _ max _ word to perform word segmentation, and the knowledge recommendation system can also adopt other methods to perform word segmentation without limitation.

In an alternative embodiment, the knowledge recommendation system may perform word segmentation on the problem to be solved to obtain a plurality of initial word segments. Then, the knowledge recommendation system can preprocess the plurality of initial participles to obtain a plurality of target participles. The pre-processing may include at least one of: stop words are removed, filter words are removed, spell correction is performed, and synonyms of the initial segmentation are added.

According to the technical scheme, the knowledge recommendation system performs word segmentation and data preprocessing on the problem to be solved, so that the knowledge recommendation system can be more accurate in determining the association score of each other knowledge and the first knowledge, and the accuracy of knowledge recommendation is further improved.

Step S123, converting the plurality of target segmented words into target feature vectors.

In the embodiment of the disclosure, the method for converting a plurality of target word segments into target feature vectors by the knowledge recommendation system includes, but is not limited to tf-idf, word2vec, seq2seq, and the like.

Step S124, calculating similarity between the target feature vector and the feature vector of the title of each knowledge in the preset knowledge base.

The knowledge recommendation system can calculate the similarity between the target feature vector and the feature vector of the title of each knowledge in the preset knowledge base by adopting algorithms such as Euclidean distance or cosine distance. Here, the smaller the distance, the more similar the representation.

In step S125, the knowledge corresponding to the title with the highest similarity is used as the first knowledge.

The knowledge recommendation system may sort the titles in order of the similarity from high to low, and use the knowledge corresponding to the title with the highest similarity as the first knowledge.

In the technical scheme provided by the embodiment of the disclosure, the knowledge recommendation system can perform word segmentation on the problem to be solved, convert a plurality of target word segments into target feature vectors, and use the knowledge corresponding to the title with the highest similarity as the first knowledge. Therefore, the user can determine the similarity according to the feature vectors, feed back the best answer and improve the accuracy and efficiency of knowledge recommendation.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, as shown in fig. 13, the method may include steps S131 to S139, and steps S131 to S138 are the same as steps S121 to S128, and are not described herein again.

And step S139, recommending other titles except the title of the first knowledge in the fifth preset number of titles with the highest similarity to the terminal equipment so that the terminal equipment displays the other titles in a navigation mode.

In the embodiment of the present disclosure, the fifth preset number may be set according to an actual requirement, for example, the fifth preset number may be 5, 8, or 10, which is not limited herein.

The knowledge recommendation system may determine a fifth preset number of the knowledge with the highest similarity from the preset knowledge base, remove the title of the first knowledge from the fifth preset number of the titles to obtain other titles, and recommend the other titles to the terminal device. The terminal device presents other titles in a navigation form. Similar questions 1-4 are displayed at the bottom of the display interface, as shown in FIG. 2. In the drawings, the number of other titles shown is merely an example, and is not limited. In fig. 2, an update button is arranged on the right side of the navigation similar questions 1-4, and when the currently presented similar questions do not have the questions in which the user is interested, the user can click the update button to update the navigation similar questions.

In order to improve the efficiency of solving problems for users, the knowledge recommendation system can sequence a plurality of titles according to the sequence of the access times of the knowledge corresponding to the titles from large to small, and recommend the sequencing result to the terminal device, so that the users can obtain the interested problems from the sequencing result, and the efficiency of recommending the knowledge is improved.

In the embodiment of the present disclosure, for each problem to be solved, the knowledge recommendation system needs to calculate the similarity between each problem to be solved and the feature vector and the feature vectors of all knowledge titles in the preset knowledge base, the machine time complexity is o (n), and n is the number of knowledge in the preset knowledge base. In step S149, a fifth preset number of titles with the highest similarity can be obtained by using the big root heap, wherein the time complexity is O (n × logK), where n is the number of knowledge in the preset knowledge base, and K is the fifth preset number. It can be seen that the time complexity of the whole process is O (n × logK), which is very large, and in order to quickly obtain the fifth preset number of titles, the knowledge recommendation system may perform dimensionality reduction on the feature vectors by using the SimHash, and obtain the fifth preset number of titles with the highest similarity by using the index, so as to reduce the time complexity of the calculation.

In the technical scheme provided by the embodiment of the disclosure, the knowledge recommendation system recommends other titles, except the title of the first knowledge, of the fifth preset number of titles with the highest similarity to the terminal device, and the titles are displayed by the terminal device. Under the navigation knowledge recommendation scene, the user can recommend the window according to the navigation knowledge, so that the problem to be solved is solved, and the problem solving efficiency of the user is improved.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, as shown in fig. 14, the method may include steps S141 to S147, and steps S141 to S142 and steps S146 to S147 are the same as steps S11 to S12 and steps S14 to S15, and are not described herein again. Steps S143-S145 are one implementation of step S13.

Step S143, mapping the theme of the first knowledge to a preset multi-dimensional feature space to obtain a first feature vector of the first knowledge.

In the embodiment of the present disclosure, after determining the first knowledge, the knowledge recommendation system may determine the association score of the first knowledge and other knowledge based on the similarity of the topic feature vector, and specifically may include steps S143 to S145.

A multi-dimensional feature space, such as an n-dimensional feature space, is preset in the knowledge recommendation system, and each dimension in the multi-dimensional feature space is corresponding to one theme. One knowledge has one or more topics. The knowledge recommendation system maps all subjects of the first knowledge to a preset multi-dimensional feature space to obtain a first feature vector of the first knowledge. The mapping rules can be set according to actual requirements.

In the embodiment of the present disclosure, the topic in the knowledge recommendation system may be a classification catalog set based on business experience, and the knowledge recommendation system may further determine the topic based on an index such as a property or MPI-score.

Step S144, mapping the theme of each other knowledge in the preset knowledge base to a preset multidimensional feature space to obtain a second feature vector of each other knowledge. Similar to step S143, reference may be made to the related description of step S143.

In the embodiment of the present disclosure, the execution order of steps S143 and S144 is not limited.

Step S145, calculating cosine similarity between the first feature vector and each second feature vector to obtain an association score between the first knowledge and each other knowledge.

In an embodiment of the present disclosure, the knowledge recommendation system may determine the association score of the first knowledge with each of the other knowledge using the following formula:

wherein score (A, B) represents the association score of knowledge A and knowledge B, n is the dimension of a preset dimensional feature space, A_iI-dimensional eigenvalues in eigenvectors representing knowledge A, B_iTo representAnd the feature value of i dimension in the feature vector of the knowledge B.

In the technical scheme provided by the embodiment of the disclosure, the knowledge recommendation system determines cosine similarity according to the subject of knowledge, further determines association score, and obtains second knowledge associated with the first knowledge, and in the process, the relevant knowledge does not need to be configured manually, so that the manpower consumption is reduced, and further, the knowledge recommendation system can accurately compare the first knowledge with all knowledge in a preset knowledge base to determine the second knowledge associated with the first knowledge, further accurately recommend relevant problems, and improve the comprehensiveness of the recommendation relevant problems, because the relevant knowledge does not need to be configured manually, even if the magnitude of knowledge stored in a knowledge base is large.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, as shown in fig. 15, the method may include steps S151 to S155, and step S151, step S152, step S154, and step S155 are the same as step S11, step S12, step S14, and step S15, and are not described herein again. Step S153 is one implementation of step S13.

Step S153, determining the same number of topics of the first knowledge and each other knowledge in the preset knowledge base as the association score of the first knowledge and the other knowledge.

In the embodiment of the present disclosure, after determining the first knowledge, the knowledge recommendation system may perform link prediction based on the local structural similarity to determine the association score of the first knowledge and other knowledge, and specifically may include step S153.

In step S153, the knowledge recommendation system counts the topics included in the first knowledge and the topics included in each of the other knowledge in the preset knowledge base; for each other knowledge, the knowledge recommendation system compares the topic included in the first knowledge with the topic included in the other knowledge, obtains the same topic of the first knowledge and the other knowledge, and further obtains the number of the same topics of the same topic of the first knowledge and the other knowledge, where the number of the same topics can be used as the association score of the first knowledge and the other knowledge.

In an optional embodiment, the knowledge recommendation system may use knowledge or a topic as a node to construct an abnormal graph, where in the abnormal graph, no edge exists between knowledge nodes, and an edge exists between a knowledge node and a topic node, and if the knowledge Q includes a topic T, an edge exists between a knowledge Q node and a topic T node; the two nodes connected by the edge are adjacent nodes. The heteromorphic graph shown in fig. 16 comprises knowledge 1-4 nodes and theme 1-4 nodes, wherein no connection exists between the knowledge 1-4 nodes, and a connection exists between the knowledge 1-4 nodes and the theme 1-4 nodes.

The knowledge recommendation system determines a neighbor node (i.e., a neighbor topic node) of the first knowledge node and determines a neighbor node of each other knowledge node based on the heteromorphic graph, and for each other knowledge node, the knowledge recommendation system may determine the same neighbor node that the first knowledge node and the other knowledge node have, where the number of the same neighbor node is the same topic number of the same topic that the first knowledge and the other knowledge have, that is, the association score of the first knowledge and the other knowledge.

The knowledge recommendation system can determine the number of the same neighbor nodes of two knowledge nodes by adopting the following formula.

score(A,B)＝|N(A)∩N(B)|。

Wherein score (a, B) represents the number of the same neighbor nodes of node a and node B, i.e. the association scores of node a and node B, n (a) represents the set of neighbor nodes of node a, and n (B) represents the set of neighbor nodes of node B.

In the technical scheme provided by the embodiment of the disclosure, based on the principle of the abnormal composition, the number of the same subjects possessed by the first knowledge and the other knowledge is determined, and the larger the number of the same subjects is, the larger the probability that an edge exists between the first knowledge and the other knowledge is, and the higher the association score between the first knowledge and the other knowledge is. Compared with a method for determining the association score based on the complex cosine similarity, the method for determining the association score based on the principle of the heteromorphic graph improves the efficiency of the association score and further improves the efficiency of knowledge recommendation.

In an embodiment of the present disclosure, a knowledge recommendation method is further provided, as shown in fig. 17, the method may include steps S171 to S176, and steps S171, S172, S175, and S176 are the same as steps S11, S12, S14, and S15, and are not repeated here. Steps S173-S174 are one possible implementation of step S13.

In step S173, a plurality of same topics of the first knowledge and each of the other knowledge in the predetermined knowledge base are determined. The manner of determining the same subject can be found in the related description of the above-mentioned step S163, and is not repeated herein.

Step S174, calculating the sum of the influence coefficients of a plurality of same subjects as the association score of the first knowledge and each other knowledge; wherein, the influence coefficient of the same theme is: the inverse of the number of knowledge having the same topic, or the inverse of the logarithm of the number of knowledge having the same topic.

For each other knowledge, for each same topic that the first knowledge and the other knowledge have, the knowledge recommendation system may calculate a reciprocal of the number of the knowledge having the same topic as an influence coefficient of the same topic corresponding to the other knowledge; the knowledge recommendation system calculates the sum of the influence coefficients of the same subject corresponding to the other knowledge as the association score of the first knowledge and the other knowledge.

The knowledge recommendation system can calculate the sum of the influence coefficients of the same subject corresponding to other knowledge by using the following formula.

Wherein score (a, B) represents influence coefficients and values of the same topic possessed by the knowledge a node and the knowledge B node point, i.e. the association score of the knowledge a node and the knowledge B node, n (a) represents a neighbor node set of the knowledge a node, n (B) represents a neighbor node set of the knowledge B node, z represents the same topic node possessed by the knowledge a node and the knowledge B node, k (z) represents the number of knowledge nodes with topic z, and k (z) in the heterogeneous graph can also be referred to as the degree of the topic z node.

For each other knowledge, for each same topic that the first knowledge and the other knowledge have, the knowledge recommendation system may calculate the inverse of the logarithm of the number of the knowledge having the same topic as the influence coefficient of the same topic corresponding to the other knowledge; the knowledge recommendation system calculates the sum of the influence coefficients of the same subject corresponding to the other knowledge as the association score of the first knowledge and the other knowledge.

Wherein score (a, B) represents influence coefficients and values of the same topic possessed by the knowledge a node and the knowledge B node, that is, association scores of the knowledge a node and the knowledge B node, n (a) represents a set of neighbor nodes of the knowledge a node, n (B) represents a set of neighbor nodes of the knowledge B node, z represents the same topic node possessed by the knowledge a node and the knowledge B node, k (z) represents the number of knowledge nodes with topic z, and k (z) may also be referred to as the degree of topic z node in the heterogeneous graph.

In the technical scheme provided by the embodiment of the disclosure, based on the principle of the heteromorphic graph, the same number of topics possessed by the first knowledge and the other knowledge is determined, and as the association score of the first knowledge and the other knowledge, the larger the same number of topics is, the larger the probability that an edge exists between the first knowledge and the other knowledge is, and the higher the association score between the first knowledge and the other knowledge is. Compared with a method for determining the association score based on the complex cosine similarity, the method for determining the association score based on the principle of the heteromorphic graph improves the efficiency of the association score and further improves the efficiency of knowledge recommendation.

In addition, in the embodiment of the present disclosure, the neighbor nodes are used as media for transferring resources, and the influence of the degree on the association of the nodes is considered, and the influence of the neighbor nodes with smaller degrees on the association relationship between the knowledge nodes is larger, and the influence of the neighbor nodes with larger degrees on the association relationship between the knowledge nodes is smaller. The accuracy and the reasonability of the association scoring are effectively improved, and the accuracy and the reasonability of knowledge recommendation are further improved.

The technical scheme provided by the embodiment of the disclosure is applicable to various scenes, such as a popular knowledge recommendation scene (shown in fig. 3), an input association recommendation scene (shown in fig. 5 and 7-9), a navigation knowledge recommendation scene (shown in fig. 12-13), an associated knowledge recommendation scene (shown in fig. 14, 15 and 17), and the like, which greatly enriches the application scenes of knowledge recommendation.

Corresponding to the above knowledge recommendation method, an embodiment of the present disclosure further provides a knowledge recommendation apparatus, as shown in fig. 18, which may include:

a first obtaining unit 181, configured to obtain a problem to be solved submitted by a terminal device;

a second acquiring unit 182 configured to acquire first knowledge for solving the problem to be solved from a preset knowledge base;

a first determining unit 183 configured to determine an association score of each of the other knowledge with the first knowledge based on the topic of the first knowledge and the topics of the other knowledge in the preset knowledge base;

a second determining unit 184, configured to determine, according to the association score, second knowledge associated with the first knowledge from the other knowledge;

the recommending unit 185 is configured to recommend the title of the second knowledge and the first knowledge to the terminal device, where the first knowledge is an answer to the problem to be solved, and the title of the second knowledge is a problem related to the problem to be solved.

Optionally, the first obtaining unit 181 may be specifically configured to:

according to the preset weight of each problem type, obtaining the problem of the problem type from the problem pool of the problem type as a hot problem;

recommending the obtained hot problems to the terminal equipment so that the terminal equipment determines a target hot problem from the hot problems;

and receiving the target hot problem determined by the terminal equipment as the problem to be solved.

Optionally, the question types include at least one of: a type of the first high frequency question, a type of the second high frequency question and a type of the preset question;

the first high frequency problem is: titles of a first preset number of knowledge with the highest access times within a preset time length;

the second high frequency problem is: and in the knowledge associated with the knowledge recommended to the terminal equipment, titles of a second preset number of knowledge with the highest access times in a preset time length are obtained.

Optionally, the first obtaining unit 181 may include:

the first acquisition subunit is used for acquiring a candidate text submitted by the terminal equipment;

the second acquisition subunit is used for acquiring a plurality of candidate titles matched with the candidate texts from the titles of the knowledge included in the preset knowledge base;

the calculating subunit is used for calculating the similarity score of each candidate title and the candidate text;

the recommendation subunit is configured to recommend a third preset number of candidate titles with the highest similarity scores to the terminal device, so that the terminal device determines a target association problem from the third preset number of candidate titles;

and the receiving subunit is used for receiving the target association problem determined by the terminal equipment as the problem to be solved.

Optionally, the second obtaining subunit may include:

performing word segmentation on the candidate text to obtain a plurality of candidate word segments;

and acquiring a title comprising the candidate participle as a candidate title from the titles of the knowledge included in the preset knowledge base.

Optionally, the calculating subunit may be specifically configured to:

calculating the ratio of a first coefficient of each candidate participle related to each candidate title to a second coefficient of the candidate participle, wherein the first coefficient of each candidate participle is the occurrence frequency of the candidate participle in the candidate title, and the second coefficient of each candidate participle is the number of the candidate titles comprising the candidate participle;

and for each candidate title, taking the sum of the ratios of all candidate participles associated with the candidate title as the similarity score of the candidate title and the candidate text.

Optionally, if there are multiple query algorithms for obtaining candidate titles, the calculating subunit may be specifically configured to:

aiming at each query algorithm, calculating the initial similarity score of each candidate title and each candidate text obtained by adopting the query algorithm;

for each candidate title, calculating the sum of the initial similarity scores of the candidate title as the similarity score of the candidate title and the candidate text.

Optionally, the first obtaining unit 181 may specifically be configured to:

recommending a plurality of hot catalogs to the terminal equipment so as to enable the terminal equipment to display the hot catalogs in a navigation mode, and determining a target catalog from the hot catalogs, wherein the hot catalogs are the catalogs which comprise a plurality of knowledge and have the access times and values higher than a preset access threshold value;

receiving a target directory submitted by terminal equipment;

recommending a fourth preset number of knowledge titles with the highest access times under the target directory to the terminal equipment so that the terminal equipment determines the target directory problem from the titles under the target directory;

and receiving the target directory problem determined by the terminal equipment as the problem to be solved.

Optionally, the second obtaining unit 182 may include:

the word segmentation sub-unit is used for segmenting words of the problem to be solved to obtain a plurality of target segmented words;

the conversion subunit is used for converting the plurality of target word segments into target feature vectors;

the calculating subunit is used for calculating the similarity between the target characteristic vector and the characteristic vector of the title of each knowledge in the preset knowledge base;

and the determining subunit is used for taking the knowledge corresponding to the title with the highest similarity as the first knowledge.

Optionally, the word segmentation subunit may be specifically configured to:

performing word segmentation on the problem to be solved to obtain a plurality of initial word segments;

preprocessing the plurality of initial participles to obtain a plurality of target participles;

the pre-processing comprises at least one of the following operations: removing stop words, removing filter words, correcting spelling, and adding synonyms of the initial participle.

Optionally, the knowledge recommendation apparatus may further include:

recommending other titles except the title of the first knowledge in a fifth preset number of titles with highest similarity to the terminal equipment so that the terminal equipment displays the other titles in a navigation mode.

Optionally, the first determining unit 183 may be specifically configured to:

mapping the theme of the first knowledge to a preset multi-dimensional feature space to obtain a first feature vector of the first knowledge;

mapping the theme of each other knowledge in a preset knowledge base to a preset multidimensional feature space to obtain a second feature vector of each other knowledge;

and calculating the cosine similarity of the first feature vector and each second feature vector to obtain the association score of the first knowledge and each other knowledge.

Optionally, the first determining unit 183 may be specifically configured to:

and determining the number of the same subjects of the first knowledge and each other knowledge in the preset knowledge base as the association score of the first knowledge and the other knowledge.

Optionally, the first determining unit 183 may be specifically configured to:

determining a plurality of same subjects of the first knowledge and each other knowledge in a preset knowledge base;

calculating the sum of the influence coefficients of a plurality of same subjects as the association score of the first knowledge and the other knowledge;

wherein, the influence coefficient of the same subject is as follows: the reciprocal of the number of knowledge having the same topic, or the reciprocal of the logarithm of the number of knowledge having the same topic.

Optionally, the second determining unit 184 may specifically be configured to:

and selecting a sixth preset number of other knowledge with highest association scores from the other knowledge as the second knowledge associated with the first knowledge.

In addition, the knowledge recommendation system can determine the second knowledge associated with the first knowledge according to the association score to further determine the related problems, the related knowledge does not need to be configured manually, manpower consumption is reduced, and the knowledge recommendation system can accurately compare the first knowledge with all knowledge in the preset knowledge base to determine the second knowledge associated with the first knowledge even if the knowledge magnitude stored in the knowledge base is large, further accurately recommend the related problems, and the comprehensiveness of the recommendation related problems is improved.

In the technical scheme of the disclosure, the collection, storage, use, processing, transmission, provision, disclosure and other processing of the personal information of the related user are all in accordance with the regulations of related laws and regulations and do not violate the good customs of the public order.

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

FIG. 19 shows a block diagram of a knowledge recommendation system 1900 for implementing the knowledge recommendation method of an embodiment of the 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. Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular telephones, 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. 19, the apparatus 1900 includes a computing unit 1901 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM)1902 or a computer program loaded from a storage unit 1908 into a Random Access Memory (RAM) 1903. In the RAM 1903, various programs and data necessary for the operation of the device 1900 can be stored. The calculation unit 1901, ROM 1902, and RAM 1903 are connected to each other via a bus 1904. An input/output (I/O) interface 1905 is also connected to bus 1904.

A number of components in device 1900 are connected to I/O interface 1905, including: an input unit 1906 such as a keyboard, a mouse, and the like; an output unit 1907 such as various types of displays, speakers, and the like; a storage unit 1908 such as a magnetic disk, optical disk, or the like; and a communication unit 1909 such as a network card, modem, wireless communication transceiver, or the like. The communication unit 1909 allows the device 1900 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.

The computing unit 1901 may be a variety of general purpose and/or special purpose processing components with processing and computing capabilities. Some examples of the computation unit 1901 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various dedicated Artificial Intelligence (AI) computation chips, various computation units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and so forth. The calculation unit 1901 executes the respective methods and processes described above, such as the knowledge recommendation method. For example, in some embodiments, the knowledge recommendation method may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as storage unit 1908. In some embodiments, part or all of a computer program may be loaded and/or installed onto the device 1900 via the ROM 1902 and/or the communication unit 1909. When the computer program is loaded into the RAM 1903 and executed by the computing unit 1901, one or more steps of the knowledge recommendation method described above may be performed. Alternatively, in other embodiments, the computing unit 1901 may be configured to perform the knowledge recommendation method in any other suitable manner (e.g., by means of firmware).

FIG. 20 is a block diagram of a knowledge recommendation system for implementing a knowledge recommendation method of an embodiment of the present disclosure, including:

at least one processor 2001; and

a memory 2002 communicatively coupled to the at least one processor 2001; wherein, the first and the second end of the pipe are connected with each other,

the memory 2002 stores instructions executable by the at least one processor 2001, the instructions being executable by the at least one processor 2001 to enable the at least one processor 2001 to perform any of the knowledge recommendation methods.

The disclosed embodiments also provide a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform a knowledge recommendation method according to any one of the above.

Embodiments of the present disclosure also provide a computer program product comprising a computer program which, when executed by a processor, implements a knowledge recommendation method according to any of the above.

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), Complex 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 or sequentially or in different orders, and are not limited herein as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved.

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, depending on 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 protection scope of the present disclosure.

Claims

1. A knowledge recommendation method, comprising:

acquiring a problem to be solved submitted by a terminal device;

2. The method of claim 1, wherein the step of acquiring the problem to be solved submitted by the terminal device comprises:

recommending the obtained plurality of hot problems to terminal equipment so that the terminal equipment determines a target hot problem from the plurality of hot problems;

and receiving the target hot problem determined by the terminal equipment as a problem to be solved.

3. The method of claim 2, wherein the issue type comprises at least one of: a type of first high frequency question, a type of second high frequency question, and a type of preset question;

the first high-frequency problem is as follows: titles of a first preset number of knowledge with the highest access times within a preset time length;

the second high-frequency problem is as follows: and in the knowledge associated with the knowledge recommended to the terminal equipment, the titles of the knowledge with the highest access frequency in the preset time length and the second preset number are selected.

4. The method of claim 1, wherein the step of acquiring the problem to be solved submitted by the terminal device comprises:

acquiring a candidate text submitted by a terminal device;

acquiring a plurality of candidate titles matched with the candidate texts from titles of knowledge included in a preset knowledge base;

calculating a similarity score of each candidate title and the candidate text;

recommending a third preset number of candidate titles with highest similarity scores to the terminal equipment so that the terminal equipment determines a target association problem from the third preset number of candidate titles;

and receiving the target association problem determined by the terminal equipment as a problem to be solved.

5. The method of claim 4, wherein the step of obtaining a plurality of candidate titles matching the candidate texts from the titles of knowledge included in the preset knowledge base comprises:

6. The method of claim 5, wherein the step of calculating a similarity score for each of the candidate headings and the candidate text comprises:

calculating the ratio of a first coefficient of each candidate participle associated with each candidate title to a second coefficient of the candidate participle, wherein the first coefficient of each candidate participle is the occurrence frequency of the candidate participle in the candidate title, and the second coefficient of each candidate participle is the number of the candidate titles comprising the candidate participle;

7. The method of claim 4 or 5, wherein there are multiple query algorithms for obtaining candidate headings, and said step of calculating a similarity score between each of said candidate headings and said candidate text comprises:

aiming at each query algorithm, calculating the initial similarity score of each candidate title and the candidate text acquired by the query algorithm;

8. The method of claim 1, wherein the step of acquiring the problem to be solved submitted by the terminal device comprises:

recommending a plurality of popular catalogs to a terminal device so that the terminal device displays the popular catalogs in a navigation mode, and determining a target catalog from the popular catalogs, wherein the popular catalog comprises a plurality of knowledge, and the number of access times and the value of the access times are higher than a preset access threshold value;

receiving the target directory submitted by the terminal equipment;

recommending a fourth preset number of knowledge titles with the highest access times under the target directory to the terminal equipment so that the terminal equipment determines a target directory problem from the titles under the target directory;

and receiving the target directory problem determined by the terminal equipment as a problem to be solved.

9. The method of claim 1, wherein the step of obtaining the first knowledge of the problem to be solved from a preset knowledge base comprises:

performing word segmentation on the problem to be solved to obtain a plurality of target word segments;

converting the plurality of target word segments into target feature vectors;

calculating the similarity between the target characteristic vector and the characteristic vector of the title of each knowledge in a preset knowledge base;

and taking the knowledge corresponding to the title with the highest similarity as the first knowledge.

10. The method of claim 9, wherein the step of segmenting the problem to be solved to obtain a plurality of target segments comprises:

the pre-processing comprises at least one of the following operations: stop words are removed, filter words are removed, spell correction is performed, and synonyms of the initial segmentation are added.

11. The method of claim 9, further comprising:

12. The method of claim 1, wherein the step of determining a relevance score for each of the other knowledge to the first knowledge based on the topic of the first knowledge and the topics of the other knowledge in the predetermined knowledge base comprises:

mapping the theme of the first knowledge to a preset multidimensional feature space to obtain a first feature vector of the first knowledge;

mapping the theme of each other knowledge in the preset knowledge base to the preset multi-dimensional feature space to obtain a second feature vector of each other knowledge;

13. The method of claim 1, wherein the step of determining a relevance score for each of the other knowledge to the first knowledge based on the topic of the first knowledge and the topics of the other knowledge in the predetermined knowledge base comprises:

14. The method of claim 1, wherein the step of determining a relevance score for each of the other knowledge to the first knowledge based on the topic of the first knowledge and the topics of the other knowledge in the predetermined knowledge base comprises:

determining a plurality of same topics of the first knowledge and each other knowledge in the preset knowledge base;

calculating a sum of the influence coefficients of the plurality of same subjects as an association score of the first knowledge with each other knowledge;

wherein the influence coefficients of the same subject are: the inverse of the number of knowledge having the same topic, or the inverse of the logarithm of the number of knowledge having the same topic.

15. The method of claim 1, wherein the step of determining second knowledge of the first knowledge association from the other knowledge based on the association score comprises:

and selecting a sixth preset number of other knowledge with highest association score from the other knowledge as second knowledge associated with the first knowledge.

16. A knowledge recommendation apparatus comprising:

the second acquisition unit is used for acquiring first knowledge for solving the problem to be solved from a preset knowledge base;

17. The apparatus according to claim 16, wherein the first obtaining unit is specifically configured to:

18. The apparatus of claim 17, wherein the issue type comprises at least one of: a type of the first high frequency question, a type of the second high frequency question and a type of the preset question;

the second high-frequency problem is as follows: and in the knowledge associated with the knowledge recommended to the terminal equipment, the titles of the knowledge with the highest access times in the preset time length and the second preset number are obtained.

19. The apparatus of claim 16, wherein the first obtaining unit comprises:

the second acquisition subunit is used for acquiring a plurality of candidate titles matched with the candidate texts from titles of knowledge included in a preset knowledge base;

and the receiving subunit is used for receiving the target association problem determined by the terminal equipment as a problem to be solved.

20. The apparatus of claim 19, wherein the second acquisition subunit comprises:

21. The apparatus according to claim 20, wherein the computing subunit is specifically configured to:

22. The apparatus according to claim 19 or 20, wherein there are a plurality of query algorithms for obtaining candidate titles, said calculating subunit is configured to:

23. A knowledge recommendation system comprising:

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 to enable the at least one processor to perform the method of any one of claims 1-15.

24. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-15.

25. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-15.