Application search method, device, equipment, storage medium and program product

Abstract

The present invention discloses a method and apparatus for searching applications, a terminal device, a storage medium and a computer program product, wherein each candidate application is searched based on application query data and a quality score of each candidate application is obtained; a preset topic model is called to analyze the topic distribution of each candidate application, wherein the topic model is trained based on preset application information; a topic score of each candidate application is calculated based on the topic distribution, and an application ranking of each candidate application is determined based on the quality score and the topic score; and each candidate application is output according to the application ranking. The present invention can achieve the arrangement of applications with a higher degree of matching with the query data input by the user, thereby effectively improving the accuracy and efficiency of application search.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and device for searching an application program, a terminal device, a storage medium, and a computer program product. Background Art

[0002] With the rapid development of smart technology, there are more and more application products used in smart terminals. For example, there are a large number of application products in the online application markets of many platforms for users to choose and download.

[0003] Since there are so many application products for users to choose from, it is very necessary to design and develop a method for querying and searching application products so that users can accurately search for the applications they want. The existing mainstream application search method usually sorts and outputs multiple candidate applications after matching them based on keywords, so that users can search again and confirm downloading in the multiple candidate applications. However, since the description of application products in the application market is usually informal and unstructured, there are still a lot of candidate applications that can be searched based on keywords in a large number of applications. Users need to spend a lot of time to search manually, and there are often no applications that users actually want to search for in the large number of candidate applications obtained by keyword matching.

[0004] In summary, the existing application search method is based on output only after keyword matching, which requires users to spend time to further search manually, resulting in low efficiency of application search. Summary of the invention

[0005] The main purpose of the present invention is to provide an application search method, device, terminal device and storage medium, aiming to solve the technical problem that the application search method in the prior art is only based on keyword matching and then output, which requires users to spend a lot of time to further search manually, resulting in low application search efficiency.

[0006] To achieve the above object, the present invention provides a method for searching for an application program, the method comprising:

[0007] Searching for each candidate application according to the application query data and obtaining a quality score of each candidate application;

[0008] Calling a preset topic model to analyze the topic distribution of each of the selected applications, wherein the topic model is trained based on preset application program information;

[0009] Calculating a theme score for each of the candidate applications according to the theme distribution, and determining an application ranking for each of the candidate applications according to the quality score and the theme score;

[0010] The candidate applications are outputted according to the order of the applications.

[0011] Furthermore, before the step of searching for each candidate application according to the application query data and obtaining a quality score of each candidate application, the method further includes:

[0012] Receive the query voice and call a preset voice recognition system to recognize the query voice to obtain application query data.

[0013] Furthermore, the preset application information includes: title information, category information and description information of the application, and the application search method further includes:

[0014] A preset crawler program is called to obtain the title information, category information and description information of the application, and the title information, category information and description information are used to train a topic model for analyzing the topic distribution of the application.

[0015] Furthermore, the application search method further includes:

[0016] A tuning operation is performed on the speech recognition system according to the title information, the category information and the description information.

[0017] Furthermore, the step of calculating the theme score of each of the selected applications according to the theme distribution includes:

[0018] Determining the semantic similarity between the subject of the candidate application and the application query data according to the subject distribution;

[0019] A topic score of the candidate application is calculated based on the semantic similarity.

[0020] Furthermore, the step of determining the application ranking of each of the candidate applications according to the quality score and the subject score includes:

[0021] The demand adaptation score of the candidate application is obtained by multiplying the quality score and the subject score of the candidate application by the quality weight and the subject weight respectively and then adding them together;

[0022] The application ranking of each of the candidate applications is determined according to the size relationship of each of the demand adaptation scores.

[0023] In addition, to achieve the above-mentioned purpose, the present invention further provides a search device for an application program, and the search device for an application program of the present invention comprises:

[0024] A search module, used to search for each candidate application according to the application query data and obtain a quality score of each candidate application;

[0025] A topic analysis module, used to call a preset topic model to analyze the topic distribution of each of the selected applications, wherein the topic model is trained based on preset application information;

[0026] A ranking determination module, configured to calculate a theme score of each of the candidate applications according to the theme distribution, and determine an application ranking of each of the candidate applications according to the quality score and the theme score;

[0027] An output module is used to output each of the candidate applications according to the order of the applications.

[0028] When running, each functional module of the application search device of the present invention implements the steps of the application search method as described above.

[0029] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal device, which includes: a memory, a processor, and a search program for applications stored in the memory and executable on the processor, and when the search program for applications is executed by the processor, the steps of the application search method as described above are implemented.

[0030] In addition, to achieve the above-mentioned purpose, the present invention also provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned application search method are implemented.

[0031] The application search method, apparatus, terminal device, storage medium and computer program product proposed in the present invention obtain each candidate application by searching according to application query data and obtaining the quality score of each candidate application; call a preset topic model to analyze the topic distribution of each candidate application, wherein the topic model is trained based on preset application information; calculate the topic score of each candidate application according to the topic distribution, and determine the application ranking of each candidate application according to the quality score and the topic score; and output each candidate application according to the application ranking.

[0032] When searching for applications, the present invention first searches for each candidate application from the application market based on the received application query data, and simultaneously obtains the quality score of each candidate application from the application market, then calls the topic model pre-trained with application information to analyze the topic distribution of each candidate application, and calculates the topic score of each candidate application based on the topic distribution, then uses the quality score and topic score of each candidate application to determine the application ranking of each candidate application, and finally outputs each candidate application according to the application ranking for the user to make a final selection.

[0033] Compared with the existing method of outputting an application only after searching for the application based on keyword matching, the present invention, after obtaining multiple candidate applications based on query data, further analyzes the subject distribution of each candidate application to calculate a subject score, and uses the subject score and the quality score of the candidate application to generate an application ranking for the multiple candidate applications. In this way, it is possible to arrange the applications with the highest degree of matching with the query data input by the user at the front, so that after outputting the various candidate applications according to the ranking, the user does not need to spend a lot of time manually searching for the required applications, which effectively improves the accuracy and efficiency of application search. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the structure of the hardware operation of the terminal device involved in the embodiment of the present invention;

[0035] Figure 2 It is a flowchart of an embodiment of a method for searching an application program of the present invention;

[0036] Figure 3 It is a schematic diagram of a process architecture involved in an embodiment of a search method for an application program of the present invention;

[0037] Figure 4 It is another schematic diagram of a process architecture involved in an embodiment of a search method for an application program of the present invention;

[0038] Figure 5 It is a schematic diagram of the module structure of a search device for an application program of the present invention.

[0039] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0040] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0041] like Figure 1 As shown, Figure 1 It is a structural diagram of the hardware operating environment of the terminal device involved in the embodiment of the present invention.

[0042] It should be noted that Figure 1 That is, it is a schematic diagram of the structure of the hardware operating environment of the terminal device. The terminal device in the embodiment of the present invention can be a PC, a portable computer and other terminal devices.

[0043] like Figure 1 As shown, the terminal device may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0044] Those skilled in the art will understand that Figure 1 The terminal device structure shown in the figure does not constitute a limitation on the terminal device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0045] like Figure 1 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a distributed task processing program. The operating system is a program that manages and controls the hardware and software resources of the sample terminal device, and supports the operation of the distributed task processing program and other software or programs.

[0046] exist Figure 1 In the terminal device shown, the user interface 1003 is mainly used to communicate data with each terminal; the network interface 1004 is mainly used to connect to the background server and communicate data with the background server; and the processor 1001 can be used to call the search program of the application stored in the memory 1005 and perform the following operations:

[0047] Searching for each candidate application according to the application query data and obtaining a quality score of each candidate application;

[0048] Calling a preset topic model to analyze the topic distribution of each of the selected applications, wherein the topic model is trained based on preset application program information;

[0049] Calculating a theme score for each of the candidate applications according to the theme distribution, and determining an application ranking for each of the candidate applications according to the quality score and the theme score;

[0050] The candidate applications are outputted according to the order of the applications.

[0051] Furthermore, the processor 1001 may call the application search program stored in the memory 1005, and before performing the search according to the application query data to obtain each candidate application and obtain the quality score of each candidate application, further perform the following operations:

[0052] Receive the query voice and call a preset voice recognition system to recognize the query voice to obtain application query data.

[0053] Furthermore, the preset application information includes: title information, category information and description information of the application. The processor 1001 may call a search program of the application stored in the memory 1005, and further perform the following operations:

[0054] A preset crawler program is called to obtain the title information, category information and description information of the application, and the title information, category information and description information are used to train a topic model for analyzing the topic distribution of the application.

[0055] Further, the processor 1001 may call a search program of an application stored in the memory 1005, and further perform the following operations:

[0056] A tuning operation is performed on the speech recognition system according to the title information, the category information and the description information.

[0057] Further, the processor 1001 may call a search program of an application stored in the memory 1005, and further perform the following operations:

[0058] Determining the semantic similarity between the subject of the candidate application and the application query data according to the subject distribution;

[0059] A topic score of the candidate application is calculated based on the semantic similarity.

[0060] Further, the processor 1001 may call a search program of an application stored in the memory 1005, and further perform the following operations:

[0061] The demand adaptation score of the candidate application is obtained by multiplying the quality score and the subject score of the candidate application by the quality weight and the subject weight respectively and then adding them together;

[0062] The application ranking of each of the candidate applications is determined according to the size relationship of each of the demand adaptation scores.

[0063] Based on the above structure, various embodiments of the application search method of the present invention are proposed.

[0064] Please refer to Figure 2 , Figure 2 FIG. 1 is a flow chart of a first embodiment of a method for searching for an application program of the present invention.

[0065] The embodiment of the present invention provides an embodiment of a method for searching an application program. It should be noted that, although a logical sequence is shown in the flow chart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0066] The method for searching for application programs in the embodiment of the present invention is applied to a terminal device. The terminal device in the embodiment of the present invention may be a PC, a portable computer or other terminal device, and no specific limitation is made here.

[0067] The search method of the application program in this embodiment includes:

[0068] Step S100, searching for each candidate application according to the application query data and obtaining a quality score of each candidate application;

[0069] In this embodiment, the terminal device receives application query data input by the user, and searches for multiple candidate applications that match the application query data based on the application query data. After querying multiple candidate applications, the terminal device further obtains quality scores of the candidate applications.

[0070] It should be noted that in this embodiment, the quality score is the score given in the current application market based on the comprehensive evaluation of the application by the user. Specifically, for example, as of December 2020, Huawei's application market has counted 46,000 users with a comprehensive score of "4.2" for the application "Weibo". In addition, in this embodiment, the application query data is based on receiving the query voice input by the user, and then recognizing the query voice to obtain the query text. It should be understood that based on the different design needs of actual applications, in other feasible implementations, the terminal device can of course also receive the query text entered by the user manually touching the physical or virtual keyboard through the user graphical interface output by the front end, and the search method of the application of the present invention is not limited to the specific method of receiving the query data of the application.

[0071] In a feasible embodiment, before the above step S100, the application search method of the present invention may further include:

[0072] Step A: receiving a query voice and calling a preset voice recognition system to recognize the query voice to obtain application query data.

[0073] It should be noted that, in this embodiment, the preset voice recognition system is a voice recognition system built into the local terminal device, or the preset voice recognition system is a third-party voice recognition system of other cloud platforms connected to the terminal device. In addition, before the voice recognition system is called by the terminal device to recognize the voice, it is also tuned based on the title information, category information and description information of the application transmitted by the terminal device. It should be understood that the voice recognition system involved in the embodiment of the present invention can independently perform recognition and analysis on the voice data and convert the voice data into accurate text data. The search method of the application of the present invention is not limited to the specific content of the voice recognition system converting the voice into text through recognition and analysis.

[0074] In this embodiment, the terminal device receives the user's query voice and then transmits the query voice to the voice recognition system. The voice recognition system recognizes and analyzes the query voice to convert the query voice into query text and feeds it back to the terminal device. The terminal device then uses the query text as application query data input by the user when he currently needs to search for an application.

[0075] Specifically, for example, please refer to Figure 3The process architecture of the embodiment of the present invention in the application stage is shown. When the terminal device is controlled by the user to enter the application search state, the terminal device receives the query voice - "user voice query" input by the user through the built-in microphone device. Then, the terminal device transmits the query voice - "user voice query" to the built-in or external voice recognition system that has been tuned in the training stage. The voice recognition system recognizes and analyzes the query voice - "user voice query" to translate the query voice - "user voice query" into a query text - "user text query". Then, the terminal device uses the query text - "user text query" as application query data, and Based on the built-in user search application search module, directly according to the application query data - "user text query", some application APP (Application) similar to the "user text query" is found from the application market as candidate applications, and the terminal device can generate an initial APP ranking of the multiple candidate applications according to the keyword matching program between the candidate application found by the search module and the application query data - "user text query". Finally, the terminal device crawls the APP quality information of each of the multiple candidate applications in the application market, thereby obtaining the quality score of the APP at the current moment.

[0076] In this embodiment, when the terminal device searches for applications based on user operations, it receives the query voice input by the user. In this way, after the terminal device enters the application search state, the user only needs to enter the voice content related to the query, breaking the limitation that the user must manually enter the query content, and improving the ease of operation for the user to search for applications.

[0077] Step S200, calling a preset topic model to analyze the topic distribution of each of the selected applications, wherein the topic model is trained based on preset application information;

[0078] It should be noted that, in this embodiment, the preset topic model is a model obtained by training a machine learning model using application-related information during the training phase of the terminal device before the application phase when the user searches for an application, and is capable of analyzing the distribution of topics to which the application belongs.

[0079] In this embodiment, after the terminal device searches for multiple candidate applications through the received application query data, it calls the currently trained topic model to analyze and obtain the topic distribution of each application, so as to subsequently use the topic distribution to sort each application according to the degree of relevance to the application that the user actually wants to search for.

[0080] Furthermore, in a feasible embodiment, the preset application information includes: title information, category information and description information of the application. The application search method of the present invention may also include:

[0081] Step B: calling a preset crawler program to obtain the title information, category information and description information of the application, and using the title information, category information and description information to train a topic model for analyzing the topic distribution of the application.

[0082] In this embodiment, the terminal device uses a crawler program to crawl the title information, category information and description information of any multiple candidate applications from the application market in advance, and then uses the title information, category information and description information of each application to perform machine learning training, thereby training a topic model that can accurately analyze the topic distribution of any application.

[0083] It should be noted that, in this embodiment, the preset crawler program can be any program for crawling available data; the title information, category information and description information of the application can be the name of the application, the category to which it belongs and the introduction. Specifically, for example, in the Huawei App Market, the title information of the application "WeChat" is "WeChat", the category information is "Social Communication", and the description information is "1. You can send voice, text messages, emoticons, pictures, and videos. 30M traffic can send and receive thousands of voice messages, saving power and traffic. 2. Moments, share life with friends. 3. Shake to see people nearby, there are no strangers in the world..."

[0084] Specifically, for example, in this embodiment, please refer to Figure 4The process architecture of the training phase of the embodiment of the present invention shown in the figure is that before the terminal device is controlled by the user to enter the application search state, it first enables the crawler program to crawl out the title information, category information, description information and other "APP information" of multiple candidate applications from the application market - "APP market". Then, the terminal device uses the title information, category information and description information of the multiple candidate applications as training data, and calls the built-in topic model training module to perform machine learning training using the training data, so that the model that can analyze the distribution of any application topic obtained after the machine learning training is completed is determined as the topic model. In this way, please continue to refer to Figure 3 The process architecture of the embodiment of the present invention in the application stage is shown, in which the terminal device, based on the built-in user search application search module, searches for some applications APP (Application) similar to the "user text query" from the application market according to the application query data - "user text query" as candidate applications, and after the initial APP ranking of the multiple candidate applications has been generated, the terminal device will further call the currently trained topic model to analyze the multiple candidate applications respectively to obtain the topic distribution of each candidate application.

[0085] Furthermore, in another feasible embodiment, the application search method of the present invention may further include:

[0086] Step C: performing a tuning operation on the speech recognition system according to the title information, category information and the description information.

[0087] In this embodiment, while the terminal device uses the title information, category information and description information of multiple candidate applications crawled from the application market to perform machine learning training to obtain a topic model, it synchronously or asynchronously tunes the speech recognition system used to recognize user input speech based on the title information, category information and description information of the multiple candidate applications.

[0088] It should be noted that in this embodiment, the tuning operation is to adjust the probability of the speech recognition system recognizing and analyzing the corresponding text based on the theme of the application. Specifically, for example, if the theme of the application is "technology", the probability of the speech recognition system recognizing "computer" and "Internet" is increased, or if the theme of the application is "fruit", the probability of the speech recognition system recognizing "orange" is increased.

[0089] Specifically, for example, please refer to Figure 4In the process architecture of the training phase of the embodiment of the present invention shown, the terminal device uses the "APP information" such as title information, category information and description information of each of the multiple candidate applications crawled from the application market to call the built-in topic model training module to use the training data to perform machine learning training to obtain the topic model. At the same time, the terminal device synchronously calls the built-in speech recognition system tuning module, and also uses the "APP information" such as title information, category information and description information of each of the multiple candidate applications to perform tuning operations on the speech recognition system used to recognize user voices to obtain query texts, thereby obtaining a tuned speech recognition system.

[0090] In this embodiment, the speech recognition system for recognizing user voice to obtain query text is optimized by utilizing the application's title information, category information, and description information, so that the speech recognition system recognizes the user's input voice to translate more accurate query text, further improving the accuracy of the user's search for applications through voice input.

[0091] Step S300, calculating the theme score of each of the candidate applications according to the theme distribution, and determining the application ranking of each of the candidate applications according to the quality score and the theme score;

[0092] In this embodiment, after the terminal device obtains the respective topic distribution of the multiple candidate applications by calling the topic model analysis, it first calculates the respective topic score of each candidate application based on the respective topic distribution of each candidate application, and then uses the respective topic score of each candidate application and the respective quality scores of the multiple candidate applications that have been crawled when searching for the multiple candidate applications through the received application query data to calculate and determine the final application ranking of each application.

[0093] Further, in a feasible embodiment, in the above step S300, the step of "calculating the theme score of each of the candidate applications according to the theme distribution" may include:

[0094] Step S301, determining the semantic similarity between the subject of the candidate application and the application query data according to the subject distribution;

[0095] Step S302: Calculate the subject score of the candidate application based on the semantic similarity.

[0096] In this embodiment, when the terminal device calculates the topic score of the candidate application based on the topic distribution of the candidate application, it first detects and determines the semantic similarity between the topic of the application identified by the topic distribution of the application and the application query data input by the user. Then, the terminal device calculates the topic score according to the semantic similarity and obtains the topic score that is positively correlated with the semantic similarity.

[0097] Specifically, for example, please refer to Figure 3 The process architecture of the embodiment of the present invention in the application stage is shown. After the terminal device calls the currently trained topic model to analyze the multiple candidate applications found respectively, so as to obtain the topic distribution of each candidate application, the terminal device will then detect the semantic similarity between the topic (such as "technology" or "fruit", etc.) identified by the topic distribution of the candidate application and the "user text query" recognized in advance by the speech recognition system after the tuning operation and used as the application query data for each candidate application. Then, after detecting the semantic similarity between the topic of the candidate application and the "user text query", the terminal device directly multiplies the semantic similarity by a predetermined score indicator to obtain To each candidate application's respective topic score, and the topic score is positively correlated with the semantic similarity between the topic of the respective application and the "user text query", that is, assuming that the terminal device detects that the semantic similarity between the topic of the current candidate application and the "user text query" is 80%, then the terminal device uses the semantic similarity - 80% multiplied by a predetermined score index (assuming it is 5) to obtain a topic score of the current candidate application of 4 points, or, assuming that the terminal device detects that the semantic similarity between the topic of another candidate application and the "user text query" is 90%, then the terminal device uses the semantic similarity - 90% multiplied by the predetermined score index to obtain a topic score of the current candidate application of 4.5 points.

[0098] Further, in a feasible embodiment, in the above step S300, the step of "determining the application ranking of each of the candidate applications according to the quality score and the subject score" may include:

[0099] Step S303, obtaining a demand adaptation score of the candidate application based on multiplying the quality score and the theme score of the candidate application by a quality weight and a theme weight respectively and then adding them together;

[0100] It should be noted that, in this embodiment, the quality weight and the subject weight are weight values ​​independently defined by business development staff based on different design requirements of actual applications. Specifically, for example, the quality weight and the subject weight can be set to 50% respectively.

[0101] Step S304: determining the application ranking of each of the candidate applications according to the magnitude relationship of each of the demand adaptation scores.

[0102] In this embodiment, after calculating the subject score of each candidate application, the terminal device further extracts the subject score of the candidate application and the quality score of the candidate application that was crawled in advance for each candidate application, and then multiplies the subject score by the subject weight and the quality score by the quality weight, and then superimposes the two products to obtain the demand adaptation score of each candidate application. Then, based on comparing the size relationship between each demand adaptation score, the terminal device rearranges each candidate application from large to small according to the size relationship to generate the final application ranking of the multiple candidate applications.

[0103] Specifically, for example, please refer to Figure 3 In the process architecture of the embodiment of the present invention in the application stage, the terminal device crawls the APP quality information of the multiple candidate applications that have been searched in the application market, thereby crawling the quality scores of the multiple candidate applications at the current moment, and using the semantic similarity between the topics identified by the topic distributions of the multiple candidate applications and the "user text query" to calculate the topic scores of the multiple candidate applications, the terminal device further extracts the quality score and topic score of each candidate application in the multiple candidate applications, and then multiplies the quality score by the quality weight to obtain The first product is obtained, and the subject score is multiplied by the subject score to obtain a second product, and then the first product and the second product are superimposed to obtain the demand adaptation score of the current candidate application. That is, assuming that the terminal device currently extracts the candidate application "Weibo" from multiple candidate applications with a quality score of 4.2 and a subject score of 4.0, the terminal device multiplies the quality score of 4.2 by the quality weight of 50% to obtain a first product of 2.1, and multiplies the subject score by the subject weight of 50% to obtain a second product of 2.0, and then the first product 2.1 and the second product 2.0 are superimposed to obtain the demand adaptation score of 4.1 for the current candidate application.

[0104] Finally, after calculating the demand adaptation score of each of the multiple candidate applications, the terminal device compares the demand adaptation scores of each of the candidate applications in turn, and arranges the candidate application with a larger demand adaptation score in front and the candidate application with a smaller demand adaptation score in the back, thereby regenerating the final application ranking of the multiple candidate applications—the final APP ranking.

[0105] Step S400: outputting the candidate applications according to the order of the applications.

[0106] In this embodiment, after the terminal device uses the subject scores and quality scores of each of the multiple candidate applications to calculate and determine the final application ranking of the multiple candidate applications, the terminal device outputs the download and installation links of the multiple candidate applications according to the application ranking, so that the user can intuitively and accurately select the application he wants to search and download and install according to the application ranking.

[0107] Specifically, for example, the terminal device finally sorts the multiple candidate applications according to APP in the form of a list, so that the candidate applications with larger adaptation scores are arranged at the front, and the multiple candidate applications are output through the front end on the user graphical interface of the user controlling the terminal device to search for applications, so that the user can accurately select the application he wants to search for from the front few candidate applications among the multiple candidate applications for downloading and installation and other operations.

[0108] In the process of the user controlling the terminal device to search for applications, the terminal device receives application query data input by the user, and searches for multiple candidate applications matching the application query data according to the application query data, and after querying the multiple candidate applications, the terminal device further obtains the quality scores of the candidate applications; after the terminal device searches for multiple candidate applications through the received application query data, it calls the currently trained topic model to analyze and obtain the respective topic distribution of each application, so as to subsequently use the topic distribution to sort each application according to the degree of relevance to the application that the user actually wants to search; after the terminal device obtains the respective topic distribution of the multiple candidate applications by calling the topic model to analyze, First, a topic score for each candidate application is calculated based on its respective topic distribution. Then, the topic score for each candidate application is used together with the quality scores of the multiple candidate applications that have been crawled when searching for multiple candidate applications through the received application query data, and a final application ranking for each application is calculated together. After the terminal device uses the topic scores and quality scores of the multiple candidate applications to calculate and determine the final application ranking for the multiple candidate applications, the terminal device outputs the download and installation links of the multiple candidate applications according to the application ranking, so that the user can intuitively and accurately select the application he wants to search for and download and install according to the application ranking.

[0109] Compared with the existing method of outputting an application only after searching for the application based on keyword matching, the present invention, after obtaining multiple candidate applications based on query data, further analyzes the subject distribution of each candidate application to calculate a subject score, and uses the subject score and the quality score of the candidate application to generate an application ranking for the multiple candidate applications. In this way, it is possible to arrange the applications with the highest degree of matching with the query data input by the user at the front, so that after outputting the various candidate applications according to the ranking, the user does not need to spend a lot of time manually searching for the required applications, which effectively improves the accuracy and efficiency of application search.

[0110] In addition, please refer to Figure 5 The embodiment of the present invention further provides a device for searching an application program, the device for searching an application program comprising:

[0111] A search module, used to search for each candidate application according to the application query data and obtain a quality score of each candidate application;

[0112] A topic analysis module, used to call a preset topic model to analyze the topic distribution of each of the selected applications, wherein the topic model is trained based on preset application information;

[0113] A ranking determination module, configured to calculate a theme score of each of the candidate applications according to the theme distribution, and determine an application ranking of each of the candidate applications according to the quality score and the theme score;

[0114] An output module is used to output each of the candidate applications according to the order of the applications.

[0115] Preferably, the application program search device further comprises:

[0116] The speech recognition module is used to receive the query speech and call a preset speech recognition system to recognize the query speech to obtain application query data.

[0117] Preferably, the preset application information includes: title information, category information and description information of the application, and the application search device further includes:

[0118] The model training module is used to call a preset crawler program to obtain the title information, category information and description information of the application, and use the title information, category information and description information to train a topic model for analyzing the topic distribution of the application.

[0119] Preferably, the application program search device further comprises:

[0120] A speech recognition tuning module is used to perform a tuning operation on the speech recognition system according to the title information, category information and the description information.

[0121] Preferably, the ranking determination module includes:

[0122] a determining unit, configured to determine, according to the topic distribution, a semantic similarity between the topic of the candidate application and the application query data;

[0123] The first calculation unit is used to calculate the topic score of the candidate application based on the semantic similarity.

[0124] Preferably, the ranking determination module further includes:

[0125] A second calculation unit, configured to obtain a demand adaptation score of the candidate application based on multiplying the quality score and the theme score of the candidate application by a quality weight and a theme weight respectively and then adding the sums;

[0126] The sorting unit is used to determine the application sorting of each of the candidate applications according to the size relationship of each of the demand adaptation scores.

[0127] The steps implemented by the functional modules of the search device of the application of the present invention during operation can refer to the various embodiments of the search method of the application of the present invention described above, and will not be described in detail here.

[0128] In addition, an embodiment of the present invention also proposes a terminal device, which includes: a memory, a processor, and a search program for an application stored in the memory and executable on the processor, wherein the search program for the application, when executed by the processor, implements the steps of the application search method as described above.

[0129] Among them, the steps implemented when the search program of the application running on the processor is executed can refer to the various embodiments of the application search method of the present invention, and will not be repeated here.

[0130] In addition, an embodiment of the present invention also proposes a storage medium, which is applied to a computer. The storage medium can be a non-volatile computer-readable storage medium, which stores a search program for an application. When the search program for the application is executed by a processor, the steps of the application search method as described above are implemented.

[0131] Among them, the steps implemented when the search program of the application running on the processor is executed can refer to the various embodiments of the application search method of the present invention, and will not be repeated here.

[0132] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0133] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0134] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0135] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for searching an application, characterized in that: The application search method includes: Searching for each candidate application according to the application query data and obtaining a quality score of each candidate application; Calling a preset topic model to analyze the topic distribution of each of the selected applications, wherein the topic model is trained based on preset application program information; The preset application information includes: title information, category information and description information of the application, and the application search method further includes: Calling a preset crawler program to obtain the title information, category information and description information of the application, and using the title information, category information and description information to train a topic model for analyzing the topic distribution of the application, wherein the preset crawler program is any program for crawling available data, and the title information, category information and description information are the name, category and introduction of the application; Calculating a theme score for each of the candidate applications according to the theme distribution, and determining an application ranking for each of the candidate applications according to the quality score and the theme score; Outputting each of the candidate applications according to the order of the applications; The method further comprises: A tuning operation is performed on the speech recognition system according to the title information, the category information and the description information. The tuning operation is to adjust the probability of the speech recognition system recognizing and analyzing the corresponding text based on the theme of the application.

2. The method for searching for an application according to claim 1, wherein: Before the step of searching for each candidate application according to the application query data and obtaining a quality score of each candidate application, the method further includes: Receive the query voice and call a preset voice recognition system to recognize the query voice to obtain application query data.

3. The method for searching for an application program according to claim 1, wherein: The step of calculating the theme score of each of the selected applications according to the theme distribution includes: Determining the semantic similarity between the subject of the candidate application and the application query data according to the subject distribution; A topic score of the candidate application is calculated based on the semantic similarity.

4. The method for searching for an application program according to claim 3, wherein: The step of determining the application ranking of each of the candidate applications according to the quality score and the subject score comprises: The demand adaptation score of the candidate application is obtained by multiplying the quality score and the subject score of the candidate application by the quality weight and the subject weight respectively and then adding them together; The application ranking of each of the candidate applications is determined according to the size relationship of each of the demand adaptation scores.

5. A device for searching an application, characterized in that: The search device of the application comprises: A search module, used to search for each candidate application according to the application query data and obtain a quality score of each candidate application; A topic analysis module, used to call a preset topic model to analyze the topic distribution of each of the selected applications, wherein the topic model is trained based on preset application information, wherein the preset application information includes: title information, category information and description information of the application, and the topic analysis module is specifically used to call a preset crawler program to obtain the title information, category information and description information of the application, and use the title information, category information and description information to train a topic model for analyzing the topic distribution of the application, wherein the preset crawler program is any program for crawling available data, and the title information, category information and description information are the name of the application, the category to which it belongs and a brief introduction; A ranking determination module, configured to calculate a theme score of each of the candidate applications according to the theme distribution, and determine an application ranking of each of the candidate applications according to the quality score and the theme score; An output module, used for outputting each of the candidate applications according to the order of the applications; The device also includes: A tuning operation is performed on the speech recognition system according to the title information, the category information and the description information. The tuning operation is to adjust the probability of the speech recognition system recognizing and analyzing the corresponding text based on the theme of the application.

6. A terminal device, characterized in that: The terminal device includes: a memory, a processor, and a search program for applications stored in the memory and executable on the processor, wherein the search program for applications, when executed by the processor, implements the steps of the application search method as described in any one of claims 1 to 4.

7. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for searching an application program according to any one of claims 1 to 4.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method for searching for an application program according to any one of claims 1 to 4 are implemented.

Images