Topic recommendation method, system, electronic device and storage medium

Abstract

The present application discloses a topic recommendation method, system, electronic device and storage medium, the method comprising: obtaining a topic recommendation request; determining candidate topics from a preset topic database according to recommended topic characteristics; performing similarity matching processing on the recommended topic characteristics and the candidate topic characteristics according to the recommended topic attributes; sorting all candidate topics according to the recommended topic sequence using all similarity results corresponding to the recommended topic characteristics until the similarity results of the candidate topics do not meet the preset sorting conditions, generating preliminary topic sorting results according to the sorted candidate topics, and sequentially expanding the recommended topic characteristics of the recommended topic attributes according to the recommended topic sequence to obtain a recommended characteristic range of the recommended topic attributes; sorting the candidate topics whose similarity results do not meet the preset sorting conditions according to the recommended characteristic range to obtain a topic expanded sorting result, which can effectively improve the diversity and accuracy of topic recommendations.

Description

Technical Field

[0001] The present application relates to but is not limited to the field of Internet education technology, and in particular to a topic recommendation method, system, electronic device and storage medium. Background Art

[0002] Similar question recommendation has a very wide application in the field of education from preschool to high school (K12). By recommending similar questions of a certain knowledge point, students can consolidate their memory of the knowledge point. In related technologies, the question features of the target question are extracted, and then the question features corresponding to each candidate question in the question bank are used to perform similarity matching, and finally the question with the highest feature similarity is used as the recommended question. However, the results obtained by sorting and recommending questions by a single category are unstable. For example, repeated questions are prone to appear frequently, or the content of the questions has a large deviation, which makes it difficult for users to get accurate questions to consolidate knowledge points. Therefore, how to further improve the diversity of question recommendations while ensuring the accuracy of question recommendations is a technical problem that needs to be solved urgently. Summary of the invention

[0003] The embodiments of the present application provide a topic recommendation method, system, electronic device, and storage medium, which can effectively improve the diversity and accuracy of topic recommendations.

[0004] In a first aspect, an embodiment of the present application provides a topic recommendation method, the method comprising:

[0005] Obtaining a topic recommendation request, the topic recommendation request including a plurality of recommended topic features and a recommended topic order, each of the recommended topic features corresponding to a recommended topic attribute, and the recommended topic order being characterized by a recommendation priority order of each of the recommended topic attributes;

[0006] According to the recommended question type feature, a plurality of candidate questions are determined from a preset question database, wherein the candidate questions include a plurality of candidate question type features, and the candidate questions are characterized as questions whose candidate question type features match at least one of the recommended question type features;

[0007] According to the recommended question type attributes, the recommended question type features and the candidate question type features are matched to obtain similarity results between the recommended question type features and the candidate question type features;

[0008] According to the recommended question class order, all the similarity results corresponding to the recommended question class features are sequentially used to sort all the candidate questions, until the similarity results of the candidate questions do not meet the preset sorting conditions, a preliminary sorting result of the questions is generated based on the sorted candidate questions, and the recommended question class features of the recommended question class attributes are sequentially expanded according to the recommended question class order to obtain a recommended feature range of at least one recommended question class attribute;.

[0009] According to at least one of the recommended feature ranges, the candidate topics whose similarity results do not meet the preset sorting condition are sorted to obtain a topic expansion sorting result;

[0010] The preliminary ranking result of the topic and the extended ranking result of the topic are integrated and processed to obtain the topic recommendation result. According to the topic recommendation method of the first aspect of the present application, there are at least the following beneficial effects: a plurality of candidate topics matching the recommended topic features are determined by preliminary matching in the preset topic database through the recommended topic features, and the scope of subsequent screening can be narrowed through preliminary matching, thereby improving the accuracy of topic recommendation. All candidate topics are sorted according to the recommended topic order in the topic recommendation request and all similarity results corresponding to each topic feature, until the similarity result of the candidate topic does not meet the preset sorting condition, a preliminary ranking result of the topic is generated according to the sorted candidate topics, and the recommended topic features of the recommended topic attributes are sequentially expanded according to the recommended topic order to obtain the recommended feature range of at least one recommended topic attribute, the candidate topics are first sorted according to the recommended topic order, and then those that do not meet the sorting condition are sorted again, and by setting different sorting conditions and recommended topic orders, a variety of similar topics can be obtained when the recommended topic features are the same, which can effectively improve the diversity of topics while ensuring the accuracy of topic recommendation. According to the order of recommended question types, the recommended question type features corresponding to the recommended question type attributes are expanded in sequence to obtain the recommended feature range. According to the recommended feature range, the candidate questions whose similarity results do not meet the preset sorting conditions are sorted. When the recommended question type features are similar and the recommended question type order is the same, different recommended questions can be expanded to further improve the diversity of questions.

[0011] According to some embodiments of the first aspect of the present application, the step of sequentially expanding the recommended question type features of the recommended question type attributes according to the recommended question type sequence to obtain a recommended feature range of at least one of the recommended question type attributes includes:

[0012] Determining the attributes of the topic to be expanded from all the attributes of the recommended topic categories according to the order of the recommended topic categories and the similarity results that do not meet the preset sorting conditions;

[0013] Determine a parameter deviation value according to the attributes of the topic to be expanded;

[0014] The recommended question type feature corresponding to the to-be-expanded question attribute is expanded according to the parameter deviation value to obtain a recommended feature range of the recommended question type attribute.

[0015] According to some embodiments of the first aspect of the present application, the step of sorting all the candidate questions according to the recommended question category order and using all the similarity results corresponding to the features of each recommended question category in sequence includes:

[0016] According to the preset requirements of the question types corresponding to the recommended question type attributes, the recommended question type sequence is modified to obtain a modified question type sequence;

[0017] According to the preset requirements of the question type corresponding to the recommended question type attribute, the similarity results of the recommended question type features under the recommended question type attribute are respectively corrected to obtain similarity correction results;

[0018] According to the question type correction order, all the candidate questions are sorted by using the similarity correction results of the recommended question type features under the recommended question type attributes in turn.

[0019] According to some embodiments of the first aspect of the present application, the candidate question type features are obtained by the following steps:

[0020] Obtain candidate questions from a preset question database;

[0021] Performing semantic feature recognition on the candidate topics to obtain multiple topic feature texts;

[0022] According to the preset question type attributes, the plurality of question feature texts are classified and processed to obtain candidate question type features corresponding to each preset question type attribute.

[0023] According to some embodiments of the first aspect of the present application, the preset question type attributes include multiple first-level question type attributes, each of the first-level question type attributes includes multiple second-level question type attributes; the multiple question feature texts are classified according to the preset question type attributes to obtain candidate question type features corresponding to each preset question type attribute, including:

[0024] According to the secondary question type attributes, the plurality of question feature texts are classified to obtain a plurality of question type attribute features;

[0025] According to the secondary question type attributes corresponding to each of the question type attribute features and the primary question type attributes to which it belongs, attribute coding processing is performed on each of the question type attribute features to obtain candidate question type features.

[0026] According to some embodiments of the first aspect of the present application, the preset sorting condition includes at least one of the following situations:

[0027] The similarity result of the candidate question under the recommended question attribute is the last one;

[0028] The similarity results of the plurality of candidate questions under the recommended question type attribute are the same;

[0029] The similarity result of the unranked candidate questions under the recommended question type attribute is lower than a preset similarity threshold.

[0030] According to some embodiments of the first aspect of the present application, the step of integrating the preliminary topic ranking result and the extended topic ranking result to obtain a topic recommendation result includes:

[0031] Get the historical data of the account that initiated the topic recommendation request;

[0032] Determining historical restriction characteristics based on the historical data;

[0033] Integrate the preliminary ranking results of the topics and the extended ranking results of the topics to obtain the recommendation results to be processed;

[0034] According to the historical restriction feature, the candidate topics in the to-be-processed recommendation results are filtered to obtain topic recommendation results.

[0035] In a second aspect, an embodiment of the present application provides a topic recommendation system, including:

[0036] An acquisition module is configured to acquire a topic recommendation request, wherein the topic recommendation request includes a plurality of recommended topic features and a recommended topic order, wherein each of the recommended topic features corresponds to a recommended topic attribute, and the recommended topic order is characterized by a recommendation priority order of each of the recommended topic attributes;

[0037] A screening module, based on the recommended question type feature, determines a plurality of candidate questions from a preset question database, wherein the candidate questions include a plurality of candidate question type features, and the candidate questions are characterized as questions whose candidate question type features match at least one of the recommended question type features;

[0038] A comparison module, based on the recommended question type attributes, performs similarity matching processing on the recommended question type features and the candidate question type features, and obtains similarity results between the recommended question type features and the features of each candidate question type;

[0039] A sorting module, according to the recommended question class order, sequentially using all the similarity results corresponding to each of the recommended question class features, sorts all the candidate questions until the similarity results of the candidate questions do not meet the preset sorting condition, generates a preliminary question sorting result based on the sorted candidate questions, and sequentially expands the recommended question class features of the recommended question class attributes according to the recommended question class order to obtain at least one recommended feature range of the recommended question class attributes; and sorts the candidate questions whose similarity results do not meet the preset sorting condition according to at least one recommended feature range to obtain a question expanded sorting result;

[0040] The recommendation module integrates the preliminary ranking result of the topics and the extended ranking result of the topics to obtain a topic recommendation result.

[0041] In a third aspect, an embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the question recommendation method of the first aspect described above is implemented.

[0042] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the method for recommending a topic according to the first aspect is implemented.

[0043] Other features and advantages of the present application will be described in the following description, and partly become apparent from the description, or understood by practicing the present application. The purpose and other advantages of the present application can be realized and obtained by the structures specifically pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are used to provide a further understanding of the technical solution of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the technical solution of the present invention and do not constitute a limitation on the technical solution of the present invention.

[0045] Figure 1 is a system architecture diagram of a processing system for executing a topic recommendation method provided in an embodiment of the present application;

[0046] Figure 2 is a flowchart of a topic recommendation method provided in an embodiment of the present application;

[0047] Figure 3 yes Figure 2 Schematic diagram of the specific implementation process of step S400;

[0048] Figure 4 yes Figure 2A schematic diagram of a specific implementation process of step S400 in another embodiment;

[0049] Figure 5 yes Figure 2 Schematic diagram of the specific implementation process of step S300;

[0050] Figure 6 yes Figure 5 Schematic diagram of the specific implementation process of step S330;

[0051] Figure 7 yes Figure 2 Schematic diagram of the specific implementation process of step S600;

[0052] Figure 8 It is a structural diagram of a topic recommendation system provided in an embodiment of the present application;

[0053] Fig. 9 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0055] It is understood that although the functional modules are divided in the device schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flow chart. The terms "first", "second", etc. in the specification, claims or the above drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0056] With the advancement of science and technology, there are many ways to improve students' memory of knowledge points. Among them, the similar question recommendation method has been widely used in preschool education to high school education. The similar question recommendation method mainly recommends similar questions with the same or similar knowledge points. In the process of continuously completing similar questions of the knowledge point, students can consolidate their memory of the knowledge point.

[0057] In the related technology, the question features of the target question are extracted, and then the question features are used to perform similarity matching with the question features corresponding to each alternative question in the question bank, and finally the question with the highest feature similarity is used as the recommended question. However, the results obtained by sorting and recommending questions by a single category are unstable. For example, repeated questions may appear frequently, or the content of the questions may deviate greatly, making it difficult for users to obtain accurate questions to consolidate their knowledge points.

[0058] In order to solve the problem that the recommendation of similar questions based on a single knowledge point is prone to unstable results, the present application provides a method, system, electronic device and storage medium for recommending questions, including obtaining a question recommendation request, wherein the question recommendation request includes multiple recommended question type features and a recommended question type sequence, each of the recommended question type features corresponds to a recommended question type attribute, and the recommended question type sequence is characterized as a recommendation priority order of each of the recommended question type attributes; according to the recommended question type features, determining multiple candidate questions from a preset question database, the candidate questions include multiple candidate question type features, and the candidate questions are characterized as questions that match the candidate question type features with at least one of the recommended question type features; according to the recommended question type attributes, performing similarity matching processing on the recommended question type features and the candidate question type features to obtain the recommended question type features. similarity results between the recommended question type feature and each of the candidate question type features; according to the recommended question type order, all of the similarity results corresponding to each of the recommended question type features are used in turn to sort all of the candidate questions, until the similarity results of the candidate questions do not meet the preset sorting conditions, a preliminary sorting result of questions is generated according to the sorted candidate questions, and the recommended question type features of the recommended question type attributes are expanded in turn according to the recommended question type order to obtain at least one recommended feature range of the recommended question type attributes; according to at least one of the recommended feature ranges, the candidate questions whose similarity results do not meet the preset sorting conditions are sorted to obtain an expanded sorting result of questions; the preliminary sorting result of questions and the expanded sorting result of questions are integrated to obtain a question recommendation result.

[0059] According to the scheme provided by the embodiment of the present application, a plurality of candidate questions matching the recommended question type features are determined by preliminary matching in a preset question database. The preliminary matching can narrow the scope of subsequent screening and improve the accuracy of question recommendations. All candidate questions are sorted according to the recommended question type order in the question recommendation request and all similarity results corresponding to each question type feature, until the similarity results of the candidate questions do not meet the preset sorting conditions, and a preliminary sorting result of the questions is generated according to the sorted candidate questions, and the recommended question type features of the recommended question type attributes are sequentially expanded according to the recommended question type order to obtain a recommended feature range of at least one recommended question type attribute, and the candidate questions are first sorted according to the recommended question type order, and then those that do not meet the sorting conditions are sorted again. By setting different sorting conditions and recommended question orders, multiple similar questions can be obtained when the recommended question type features are the same, which can effectively improve the diversity of questions while ensuring the accuracy of question recommendations. According to the order of recommended question types, the recommended question type features corresponding to the recommended question type attributes are expanded in sequence to obtain the recommended feature range. According to the recommended feature range, the candidate questions whose similarity results do not meet the preset sorting conditions are sorted. When the recommended question type features are similar and the recommended question type order is the same, different recommended questions can be expanded to further improve the diversity of questions.

[0060] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0061] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for the normal operation of the embodiment of the present application is obtained. The following is a further elaboration of the embodiment of the present application in conjunction with the accompanying drawings.

[0062] Reference Figure 1 A processing system 100 for executing a topic recommendation method is shown. The processing system 100 includes an application server 110 and a client 120. After obtaining a topic recommendation request sent by the client 120, the application server 110 first determines multiple candidate topics in a preset topic database according to the recommended topic class features in the topic recommendation request. Secondly, according to the recommended topic class attributes, the recommended topic class features and the candidate topic class features are matched with each other to obtain the similarity results of the recommended topic class features and each candidate topic class features. Thirdly, according to the recommended topic class order, all the candidate topics are sorted by using all the similarity results corresponding to each recommended topic class feature in turn, until the similarity results of the candidate topics do not meet the preset sorting conditions, a preliminary topic sorting result is generated according to the sorted candidate topics, and the recommended topic class features of the recommended topic class attributes are sequentially expanded according to the recommended topic class order to obtain at least one recommended feature range of the recommended topic class attributes. Finally, according to at least one recommended feature range, the candidate topics whose similarity results do not meet the preset sorting conditions are sorted to obtain the topic expansion sorting result, and the preliminary topic sorting result and the topic expansion sorting result are integrated to obtain the topic recommendation result. After obtaining the topic recommendation result, the application server 100 will push the topic recommendation result to the client 120 .

[0063] The processing system 100 for executing the question recommendation method described in the embodiment of the present application is for the purpose of more clearly illustrating the technical solution of the embodiment of the present application, and does not constitute a limitation on the technical solution provided in the embodiment of the present application. Those skilled in the art will appreciate that with the evolution of the processing system 100 and the emergence of new application scenarios, the technical solution provided in the embodiment of the present application is also applicable to similar technical problems.

[0064] It can be understood by those skilled in the art that Figure 1The structure of the processing system 100 shown in the figure does not constitute a limitation on the embodiments of the present application, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0065] Based on the structure of the processing system 100 described above, various embodiments of the topic recommendation method of the present application are proposed.

[0066] First, refer to Figure 2 , Figure 2 The flowchart of the topic recommendation method provided by the embodiment of the present application is shown, and the topic recommendation method includes but is not limited to the following steps:

[0067] Step S100, obtaining a topic recommendation request, the topic recommendation request including a plurality of recommended topic category features and a recommended topic category sequence;

[0068] Step S200, determining a plurality of candidate questions from a preset question database according to the characteristics of the recommended question type;

[0069] Step S300, performing similarity matching processing on the recommended question type features and the candidate question type features according to the recommended question type attributes, and obtaining similarity results between the recommended question type features and the features of each candidate question type;

[0070] Step S400, according to the order of recommended question categories, all similarity results corresponding to the recommended question category features are used in turn to sort all candidate questions, until the similarity results of the candidate questions do not meet the preset sorting conditions, a preliminary sorting result of the questions is generated according to the sorted candidate questions, and the recommended question category features of the recommended question category attributes are sequentially expanded according to the order of recommended question categories to obtain a recommended feature range of at least one recommended question category attribute;

[0071] Step S500, sorting the candidate topics whose similarity results do not meet the preset sorting condition according to at least one recommended feature range, to obtain a topic expansion sorting result;

[0072] Step S600: Integrate the preliminary ranking results and the extended ranking results to obtain a topic recommendation result.

[0073] It is understandable that the topic recommendation request from the client is received by the application server, and the topic recommendation request includes multiple recommended topic features and a recommended topic order, each of which corresponds to a recommended topic attribute, wherein the recommended topic feature can correspond to one recommended topic attribute or multiple recommended topic attributes, and the recommended topic order is characterized by the recommendation priority order of each recommended topic attribute. After obtaining the recommended topic features, multiple candidate topics are determined in the preset topic database, and the candidate topics include multiple candidate topic features. The candidate topics are characterized as topics that match the candidate topic features with at least one recommended topic feature, and each candidate topic feature corresponds to a candidate topic attribute. After determining the candidate topics, the recommended topic features and the candidate topic features are matched according to the recommended topic attributes to obtain the similarity results of the recommended topic features and each candidate topic feature. The higher the similarity, the higher the correlation between the candidate topic attribute of the candidate topic feature and the recommended topic attribute of the recommended topic feature. After calculating the similarity results, all candidate questions are sorted according to the order of recommended question types, using all similarity results corresponding to each recommended question type feature in turn, until the similarity structure of the candidate question does not meet the preset sorting conditions. The preliminary sorting results of the questions are generated based on the sorted candidate questions, which can ensure that each question is sorted only once, thereby improving accuracy. The preset sorting condition is to sort only the questions whose similarity results of the candidate questions are greater than the preset similarity result threshold, so that the recommended questions are more in line with the expected goals. After determining the pre-sorting results, the recommended question type features of the recommended question type attributes are expanded in turn according to the order of recommended question types to obtain the recommended feature range of the recommended question type attributes. According to the recommended feature range, the candidate questions whose similarity results do not meet the preset sorting conditions are sorted to obtain the expanded sorting results of the questions. Finally, the preliminary sorting results of the questions and the expanded sorting results of the questions are integrated to obtain the question recommendation results.

[0074] It should be noted that the order of recommended question types can be a system preset order or a user set order, wherein the user can set the recommended priority order of some recommended question type attributes and set the order of all recommended question types. When the user only sets the recommended priority order of some recommended question type attributes, the application server will complete the unselected recommended question type attributes according to the partial recommended question type attributes set by the user. For example, in another embodiment, the system presets the priority order of subject, school system, difficulty, knowledge point, grade, thinking, questioning method, pictures and texts, and question types. The user only selects difficulty as the first priority, and the system will take difficulty as the first priority, and then prioritize subject, school system, knowledge point, grade, thinking, questioning method, pictures and texts, and question types in turn. By setting the order of recommended question types by the user, the recommended questions can be more in line with the user's own needs, and the recommended questions generated by different recommendation priority orders will also be different, effectively improving the diversity of question recommendations.

[0075] It should be noted that the similarity matching process may use natural language processing technology to determine the correlation between the recommended question attributes of the recommended question features and the candidate question attributes of the candidate question features. The higher the correlation, the higher the similarity.

[0076] It is understandable that in other embodiments, the recommended question features and the candidate question features can be matched individually and weighted. When the recommended question features and the candidate question features are matched individually, each feature will be matched individually with each candidate question feature to obtain a similarity result between each recommended question feature and each candidate question feature. When the recommended question features and the candidate question features are weighted matched, after obtaining the similarity result between the recommended question features and each candidate question feature, the similarity result is weighted according to the recommended question order, for example, the weighted value is decreased in descending order according to the recommended question order, and then the similarity result is multiplied by the weighted value.

[0077] It is understandable that the preset sorting condition can also be that the similarity result of the candidate question under the recommended question type attribute is the last one, that is, when sorting, the similarity results of the last recommended question type attribute are not sorted, but the unsorted questions are sorted according to the recommended feature range of the recommended question type attribute, which can further avoid the occurrence of the same question and further improve the diversity of the recommended questions.

[0078] It is understandable that the preset sorting condition can also be that the similarity results of multiple candidate questions under the recommended question type attribute are the same. If the similarity results are the same, the questions can be considered to be the same. When the similarity results of multiple candidate questions under the recommended question type attribute are the same, the sorting of the candidate questions corresponding to the recommended question type attribute is skipped to avoid multiple repeated questions when the user is answering the questions, which affects the user's experience.

[0079] It is understandable that the recommended question type feature can be text or code. When the recommended question type feature is text, the recommended question type attribute corresponding to the recommended question type feature is also text. When the recommended question type feature is code, the recommended question type attribute corresponding to the recommended question type feature is also code. When the recommended question type feature is code, the application server will also encode the question type feature and question type attribute corresponding to the question in the preset question database before performing similarity matching.

[0080] It should be noted that different question type attributes correspond to different codes. For example, when the identified question type attribute is the first grade, the code is 001, and the code for the second grade is 002. For another example, when the identified question type attribute is a certain word, the code is 100, and the code for the word opposite to the word is 099. The word with the highest similarity to the word is 101, and the word with the second highest similarity is 102.

[0081] It should be noted that when a question type feature corresponds to multiple question type attributes, the encoding of the question type feature is composed of the encodings of the multiple question type attributes. For example, when a question type feature corresponds to two question type attributes coded as 101 and 020, the encoding of the question type feature is 101020.

[0082] Reference Figure 3 , Figure 2 Step S400 in the illustrated embodiment includes but is not limited to the following steps:

[0083] Step S410, determining the attributes of the topic to be expanded from all the attributes of the recommended topic categories according to the order of the recommended topic categories and the similarity results that do not meet the preset sorting conditions;

[0084] Step S420, determining a parameter deviation value according to the attributes of the topic to be expanded;

[0085] Step S430, expanding the recommended question type feature corresponding to the to-be-expanded question attribute according to the parameter deviation value, and obtaining a recommended feature range of the recommended question type attribute.

[0086] It is understandable that, according to the order of the recommended question types and the similarity results that do not meet the preset conditions, the attributes of the topic to be expanded are determined from all the recommended question type attributes, wherein the attributes of the topic to be expanded are the topic attributes corresponding to the recommended question types whose similarity results do not meet the preset conditions in the recommended question type characteristics. After determining the attributes of the topic to be expanded, the application server will determine the parameter deviation value of the topic to be expanded according to different attributes of the topic to be expanded. For example, when the attribute of the topic to be expanded is a certain grade, the parameter deviation value is the number of levels that the grade increases or decreases. For another example, when the attribute of the topic to be expanded is a certain knowledge point, the parameter deviation value is the word association range of the knowledge point. After determining the parameter deviation value, the recommended question type feature corresponding to the attribute of the topic to be expanded is expanded according to the parameter deviation value to obtain the recommended feature range of the recommended question type attribute. For example, if the topic to be expanded is the fourth grade, the parameter deviation value is the grade increase or decrease of 1, and the recommended question type feature range is from the third grade to the fifth grade.

[0087] It should be noted that when the question feature text is a code, the recommended question type feature corresponding to the question attribute to be expanded is expanded according to the parameter deviation value to obtain the recommended feature range of the recommended question type attribute, and the recommended feature range is the increase or decrease range of the code.

[0088] It should be noted that the parameter deviation value may also be a parameter deviation value set by a user.

[0089] Reference Figure 4 , Figure 2 Step S400 in the illustrated embodiment also includes but is not limited to the following steps:

[0090] Step S440, according to the preset requirements of the recommended question types corresponding to the recommended question type attributes, the recommended question type sequence is corrected to obtain a corrected question type sequence;

[0091] Step S450, according to the preset requirements of the recommended question type corresponding to the recommended question type attribute, the similarity results of the recommended question type features under the recommended question type attribute are respectively corrected to obtain similarity correction results;

[0092] Step S460, according to the question type correction order, all candidate questions are sorted by using the similarity correction results of the recommended question type features under the recommended question type attributes in sequence.

[0093] It is understandable that the application server will also modify the order of recommended question types according to the preset requirements of the question types corresponding to the recommended question type attributes to obtain the corrected order of question types, wherein the preset requirements of the question types can be the priority of a certain attribute or the similarity result requirements of some attributes. For example, if the preset requirements of the question types include that the priority of multiple-choice questions is the highest, the application server will set the priority of multiple-choice questions to the highest in the original recommended question type order and generate a new corrected order of question types. After similarity matching, the application server will modify the similarity results of the recommended question type features corresponding to the recommended question type attributes according to the preset requirements of the question types corresponding to the recommended question type attributes to obtain the corrected similarity results. For example, if the preset requirements of the question types include that some question type attributes must be the same and some question type attributes must be different, that is, the similarity of some different question type features is negated. Finally, according to the corrected order of question types, all candidate questions are sorted using the corrected similarity results of the recommended question type features under the recommended question type attributes. By correcting the order of recommended question categories and the similarity results of recommended question category characteristics according to the preset requirements of the question categories corresponding to the recommended question category attributes, and then sorting all candidate questions, the accuracy of the recommended questions can be further improved.

[0094] It should be noted that the question type preset requirements may be question type preset requirements pre-set by the user, or the application server may generate different question type preset requirements according to the recommended question type attributes set by the user.

[0095] Reference Figure 5 , Figure 2 The candidate question type features in step S300 of the illustrated embodiment are obtained by the following steps:

[0096] Step S310, obtaining candidate topics in a preset topic database;

[0097] Step S320, performing semantic feature recognition on the candidate topics to obtain a plurality of topic feature texts;

[0098] Step S330: classify the multiple question feature texts according to the preset question type attributes to obtain candidate question type features corresponding to each preset question type attribute.

[0099] It is understandable that before similarity matching, the application server will obtain candidate questions in the preset question database, and then perform semantic feature recognition on the candidate questions to obtain multiple question feature texts, wherein the question feature texts can be text or code. According to the preset question type attributes, the multiple question feature texts are classified and processed to obtain the candidate question type features corresponding to each preset question type attribute. By classifying the multiple question feature texts, the matching efficiency can be improved during the subsequent similarity matching process.

[0100] It should be noted that semantic feature recognition refers to a natural language processing technology that can effectively identify the meaning contained in each sentence and gain an in-depth understanding of the sentences and their contents. With the help of natural language processing technology and machine learning, this technology can parse grammar, characters, etc. from sentences based on the user's natural input, achieving more readable and accurate text understanding.

[0101] Reference Figure 6 , Figure 5 Step S330 in the illustrated embodiment includes but is not limited to the following steps:

[0102] Step S331, classifying multiple question feature texts according to the secondary question type attributes to obtain multiple question type attribute features;

[0103] Step S332, performing attribute coding processing on each question type attribute feature according to the secondary question type attribute corresponding to each question type attribute feature and the primary question type attribute to which it belongs, to obtain a candidate question type feature.

[0104] It is understandable that the application server will also classify multiple question feature texts according to the secondary question attributes to obtain multiple question attribute features, where the secondary question attributes are the attributes corresponding to the primary question attributes. Classifying the secondary question attributes can improve the matching efficiency during subsequent matching. Then, according to the secondary question attributes corresponding to each question attribute feature and the primary question attributes to which it belongs, attribute coding is performed on each question attribute feature to obtain candidate question features. By identifying and classifying the secondary question attributes, the matching parameters of the question features can be improved, thereby improving the accuracy of matching.

[0105] It is understandable that, in another embodiment, the secondary question attribute may also correspond to the tertiary question attribute, and the application server may further classify the multiple question feature texts according to the tertiary question attribute. By identifying and classifying the multi-level question attributes, the accuracy of matching can be effectively improved. For example, the question features include the primary question attributes such as words, rhetoric, and reading. The words include the secondary question attributes such as nouns, verbs, adjectives, and adverbs. The adverbs include time, frequency, range, mood, and degree adverbs.

[0106] It is understandable that the preset requirements for question types may also limit the priority or similarity requirements of secondary question type attributes, thereby modifying the ranking of candidate questions.

[0107] Reference Figure 7 , Figure 2 Step 600 in the illustrated embodiment includes but is not limited to the following steps:

[0108] Step S610, obtaining historical data of the account that initiated the topic recommendation request;

[0109] Step S620, determining historical restriction features based on historical data;

[0110] Step S630, integrating the preliminary ranking results and the extended ranking results to obtain the recommendation results to be processed;

[0111] Step S640: filter the candidate topics in the recommendation results to be processed according to the historical restriction characteristics to obtain topic recommendation results.

[0112] It is understandable that the application server obtains the historical data of the account that initiates the topic recommendation request, where the historical data includes the topics that the user has already done, the topics that the user has answered correctly, and the features or attributes that are excluded by the historical selection. Then, based on the historical data, the historical restriction features are determined. After the application server completes the integration of the preliminary ranking results of the topics and the expanded ranking results of the topics, it will filter the candidate topics in the recommended results to be processed according to the historical restriction features to obtain the topic recommendation results. By filtering the pre-selected topics, users are prevented from doing repeated topics or topics that do not meet their expectations, the diversity of the recommended topics is increased, and the training effect is thereby improved.

[0113] It should be noted that, in another embodiment, the user can also set filtering conditions. According to the filtering conditions set by the user, the recommended questions can be more in line with the user's expectations and improve the user's usage experience. For example, if the filtering condition set by the user is wrong questions, then the application server will not only recommend historical wrong questions to the user, but also perform similarity matching in the question database based on the historical wrong questions and recommend questions with higher similarity, thereby improving the user's knowledge point memory effect.

[0114] Second, refer to Figure 8 , Figure 8 A schematic diagram of the structure of a topic recommendation system 800 provided in an embodiment of the present invention is shown.

[0115] The acquisition module 810 acquires a topic recommendation request, the topic recommendation request includes a plurality of recommended topic features and a recommended topic order, each recommended topic feature corresponds to a recommended topic attribute, and the recommended topic order is represented by a recommendation priority order of each recommended topic attribute;

[0116] A screening module 820 determines a plurality of candidate questions from a preset question database according to the recommended question type feature, wherein the candidate questions include a plurality of candidate question type features, and the candidate questions are characterized as questions whose candidate question type features match at least one recommended question type feature;

[0117] Comparison module 830, based on the recommended question attributes, performs similarity matching processing on the recommended question features and the candidate question features, and obtains similarity results between the recommended question features and the features of each candidate question;

[0118] The sorting module 840 sorts all candidate questions according to the order of recommended question categories and sequentially uses all similarity results corresponding to the features of each recommended question category until the similarity results of the candidate questions do not meet the preset sorting conditions, generates preliminary sorting results of questions according to the sorted candidate questions, and sequentially expands the recommended question category features of the recommended question category attributes according to the order of recommended question categories to obtain a recommended feature range of at least one recommended question category attribute; sorts the candidate questions whose similarity results do not meet the preset sorting conditions according to the at least one recommended feature range to obtain an expanded sorting result of questions;

[0119] The recommendation module 850 integrates the preliminary topic ranking results and the extended topic ranking results to obtain topic recommendation results.

[0120] Thirdly, refer to Fig. 9 , Fig. 9 The electronic device 900 provided in an embodiment of the present application is shown. The electronic device 900 includes a memory 910, a processor 920, and a computer program stored in the memory 910 and executable on the processor 920. When the processor 920 executes the computer program, the topic recommendation method in the above embodiment is implemented.

[0121] The memory 910 is a non-transitory computer-readable storage medium that can be used to store non-transitory software programs and non-transitory computer executable programs, such as the topic recommendation method in the above embodiment of the present application. The processor 920 implements the topic recommendation method in the above embodiment of the present application by running the non-transitory software programs and instructions stored in the memory 910.

[0122] The memory 910 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data required to execute the topic recommendation method in the above embodiment, etc. In addition, the memory 910 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. It should be noted that the memory 910 may optionally include a memory remotely arranged relative to the processor 920, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0123] The non-transient software program and instructions required to implement the topic recommendation method in the above embodiment are stored in the memory, and when executed by one or more processors, the topic recommendation method in the above embodiment is executed, for example, the above described Figure 2 Steps S100 to S600 of the method, Figure 3 Steps S410 to S430 of the method, Figure 4 Steps S440 to S460 of the method, Figure 5 Steps S310 to S330 of the method, Figure 6 Steps S331 to S332 of the method and Figure 7 Method steps S610 to S640.

[0124] In a fourth aspect, the present application further provides a computer-readable storage medium, the computer-readable storage medium storing computer-executable instructions, the computer-executable instructions being used to cause a computer to execute the topic recommendation method in the above embodiment, for example, to execute the above described Figure 2 Steps S100 to S600 of the method, Figure 3 Steps S410 to S430 of the method, Figure 4 Steps S440 to S460 of the method, Figure 5 Steps S310 to S330 of the method, Figure 6 Steps S331 to S332 of the method and Figure 7 Method steps S610 to S640.

[0125] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0126] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0127] It should be noted that the server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), as well as big data and artificial intelligence platforms.

[0128] It should be noted that all or some of the steps in the method disclosed above can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0129] The embodiments of the present application are described in detail above in conjunction with the accompanying drawings, but the present application is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in the technical field without departing from the purpose of the present application.

Claims

1. A topic recommendation method, characterized in that: include: Obtaining a topic recommendation request, the topic recommendation request including a plurality of recommended topic features and a recommended topic order, each of the recommended topic features corresponding to a recommended topic attribute, and the recommended topic order being characterized by a recommendation priority order of each of the recommended topic attributes; According to the recommended question type feature, a plurality of candidate questions are determined from a preset question database, wherein the candidate questions include a plurality of candidate question type features, and the candidate questions are characterized as questions whose candidate question type features match at least one of the recommended question type features; According to the recommended question type attributes, the recommended question type features and the candidate question type features are matched to obtain similarity results between the recommended question type features and the candidate question type features; According to the recommended question class order, all the similarity results corresponding to the recommended question class features are sequentially used to sort all the candidate questions until the similarity results of the candidate questions do not meet the preset sorting conditions, a preliminary sorting result of questions is generated according to the sorted candidate questions, and the recommended question class features of the recommended question class attributes are sequentially expanded according to the recommended question class order to obtain a recommended feature range of at least one recommended question class attribute; According to at least one of the recommended feature ranges, the candidate topics whose similarity results do not meet the preset sorting condition are sorted to obtain a topic expansion sorting result; The preliminary ranking result of the topics and the extended ranking result of the topics are integrated to obtain a topic recommendation result.

2. The topic recommendation method according to claim 1, characterized in that: The step of sequentially expanding the recommended question type features of the recommended question type attributes according to the recommended question type sequence to obtain a recommended feature range of at least one of the recommended question type attributes includes: Determining the attributes of the topic to be expanded from all the attributes of the recommended topic categories according to the order of the recommended topic categories and the similarity results that do not meet the preset sorting conditions; Determine a parameter deviation value according to the attributes of the topic to be expanded; The recommended question type feature corresponding to the to-be-expanded question attribute is expanded according to the parameter deviation value to obtain a recommended feature range of the recommended question type attribute.

3. The topic recommendation method according to claim 1, characterized in that: The step of sorting all the candidate questions according to the order of the recommended question categories and using all the similarity results corresponding to the features of each recommended question category in turn includes: According to the preset requirements of the question types corresponding to the recommended question type attributes, the recommended question type sequence is modified to obtain a modified question type sequence; According to the preset requirements of the question type corresponding to the recommended question type attribute, the similarity results of the recommended question type features under the recommended question type attribute are respectively corrected to obtain similarity correction results; According to the question type correction order, all the candidate questions are sorted by using the similarity correction results of the recommended question type features under the recommended question type attributes in turn.

4. The topic recommendation method according to claim 1, characterized in that: The candidate question type features are obtained by the following steps: Obtain candidate questions from a preset question database; Performing semantic feature recognition on the candidate topics to obtain multiple topic feature texts; According to the preset question type attributes, the plurality of question feature texts are classified and processed to obtain candidate question type features corresponding to each preset question type attribute.

5. The topic recommendation method according to claim 4, characterized in that: The preset question type attributes include a plurality of first-level question type attributes, and each of the first-level question type attributes includes a plurality of second-level question type attributes; The method of classifying the plurality of question feature texts according to the preset question type attributes to obtain candidate question type features corresponding to each preset question type attribute includes: According to the secondary question type attributes, the plurality of question feature texts are classified to obtain a plurality of question type attribute features; According to the secondary question type attributes corresponding to each of the question type attribute features and the primary question type attributes to which it belongs, attribute coding processing is performed on each of the question type attribute features to obtain candidate question type features.

6. The topic recommendation method according to claim 1, characterized in that: The preset sorting condition includes at least one of the following situations: The similarity result of the candidate question under the recommended question attribute is the last one; The similarity results of the plurality of candidate questions under the recommended question type attribute are the same; The similarity result of the unranked candidate questions under the recommended question type attribute is lower than a preset similarity threshold.

7. The topic recommendation method according to claim 1, characterized in that: The step of integrating the preliminary topic ranking result and the extended topic ranking result to obtain a topic recommendation result includes: Get the historical data of the account that initiated the topic recommendation request; Determining historical restriction characteristics based on the historical data; Integrate the preliminary ranking results of the topics and the extended ranking results of the topics to obtain the recommendation results to be processed; According to the historical restriction feature, the candidate topics in the to-be-processed recommendation results are filtered to obtain topic recommendation results.

8. A topic recommendation system, characterized in that: include: An acquisition module is configured to acquire a topic recommendation request, wherein the topic recommendation request includes a plurality of recommended topic features and a recommended topic order, wherein each of the recommended topic features corresponds to a recommended topic attribute, and the recommended topic order is characterized by a recommendation priority order of each of the recommended topic attributes; A screening module, based on the recommended question type feature, determines a plurality of candidate questions from a preset question database, wherein the candidate questions include a plurality of candidate question type features, and the candidate questions are characterized as questions whose candidate question type features match at least one of the recommended question type features; A comparison module, based on the recommended question type attributes, performs similarity matching processing on the recommended question type features and the candidate question type features, and obtains similarity results between the recommended question type features and the features of each candidate question type; A sorting module, according to the recommended question class order, sequentially using all the similarity results corresponding to each of the recommended question class features, sorts all the candidate questions until the similarity results of the candidate questions do not meet the preset sorting condition, generates a preliminary question sorting result based on the sorted candidate questions, and sequentially expands the recommended question class features of the recommended question class attributes according to the recommended question class order to obtain at least one recommended feature range of the recommended question class attributes; and sorts the candidate questions whose similarity results do not meet the preset sorting condition according to at least one recommended feature range to obtain a question expanded sorting result; The recommendation module integrates the preliminary ranking result of the topics and the extended ranking result of the topics to obtain a topic recommendation result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the topic recommendation method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the topic recommendation method according to any one of claims 1 to 7 is implemented.

Images