A method and device for determining business problems

Abstract

The present invention discloses a method and device for determining business problems, and relates to the field of computer technology. A specific implementation of the method includes: performing semantic analysis on the acquired audit opinion data to obtain and save the sentence vectors of the audit opinion data; querying the sentence vectors that meet the conditions from the sentence vectors of the stored audit opinion data, and clustering the queried sentence vectors; determining the business problem to be solved according to the sentence vectors of each category obtained by clustering. This implementation can automatically find out the existing potential problem points, which can effectively reduce the cognitive bias caused by subjective reasons when manually analyzing business problems, directly hit the business pain points, avoid the waste of R&D resources, and reduce the number of information transmissions between people, thereby reducing information loss, improving the accuracy of positioning needs, and discovering potential problems before the problem actually occurs to optimize the system, so as to avoid business problems being displayed to users and affecting user experience.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for determining a business problem. Background Art

[0002] Behind every Internet product is an operation team and an operation management system. As the first line of contact with customers, the operation team is often exposed to a variety of problem feedback. The entire process of current business problems from generation to final implementation to promote product iteration goes through two important nodes. One is the summary and refinement of the product manager of the operation team, and the other is the understanding and transformation of the product manager of the technical team. It is usually completed through the thinking of each product manager plus some simple data statistics.

[0003] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0004] The determination of business problems relies on the analysis and summary of the product manager of the operation team, which is a great test of the product manager's understanding of the business and his ability to summarize and generalize. The proposed requirements are likely to consume a lot of R&D resources and ultimately fail to truly solve the business pain points, with little effect;

[0005] In the existing process, since everyone thinks differently, a message transmission may mean the loss of some information, and the product requirement document obtained by the technicians may be very different from the original requirements.

[0006] Business problems can only be identified after they have already occurred, affecting the user experience. Summary of the invention

[0007] In view of this, the embodiments of the present invention provide a method and device for determining business problems, which can automatically find out the existing potential problem points, effectively reduce the cognitive bias caused by subjective reasons when manually analyzing business problems, directly hit the business pain points, avoid waste of R&D resources, and reduce the number of information transmissions between people, thereby reducing information loss and improving the accuracy of positioning needs. Potential problems can be discovered before the problem actually occurs to optimize the system, avoiding business problems from being presented to users and affecting user experience.

[0008] To achieve the above objective, according to one aspect of an embodiment of the present invention, a method for determining a business problem is provided.

[0009] A method for determining a business problem includes: performing semantic analysis on acquired audit opinion data to obtain and save sentence vectors of the audit opinion data; querying sentence vectors that meet conditions from the stored sentence vectors of the audit opinion data, and clustering the query sentence vectors; and determining the business problem to be solved based on the sentence vectors of each category obtained by clustering.

[0010] Optionally, a scheduled task is used to regularly query and obtain new audit opinion data within a recent preset time period, and the semantic analysis is performed on each audit opinion data to obtain a sentence vector of the new audit opinion data, and save it to a database, which is used to store the sentence vectors of the stored audit opinion data.

[0011] Optionally, semantic analysis is performed on audit opinion data in the following manner: the audit opinion data is segmented to obtain a segmentation array; a word vector model is used to calculate a word vector for each word in the segmentation array, and the obtained word vectors are weighted averaged according to the weights corresponding to the parts of speech to obtain a sentence vector for the audit opinion data.

[0012] Optionally, determining the business problem to be solved based on the sentence vectors of each category obtained by clustering includes: determining the business problem indicated by the category with the largest number of sentence vectors among the sentence vectors of each category obtained by clustering as the business problem to be solved.

[0013] Optionally, determining the business problem to be solved based on the sentence vectors of each category obtained by clustering includes: respectively calculating the distance between each sentence vector of each category obtained by clustering and the center point vector of the category to which it belongs; determining the convergence degree of the sentence vectors of each category based on the distance and the number of sentence vectors of each category; and determining the business problem indicated by the category whose convergence degree meets a preset condition as the business problem to be solved.

[0014] Optionally, determining the convergence degree of the sentence vectors of each category based on the distance and the number of sentence vectors of each category includes: sorting the sentence vectors of each category according to the distance from the center point vector of the corresponding category to obtain a sentence vector sequence of each category; for each category sentence vector sequence, calculating the proportion of the number of sentence vectors in the category sentence vector sequence whose distance is less than a distance threshold in the total number of sentence vectors of the category to obtain the convergence degree of the sentence vector of the category.

[0015] Optionally, the distance threshold is determined in the following manner: forming multiple sentence sets, each sentence set including multiple sentences with similar semantics; converting the sentences in the sentence sets into sentence vectors respectively, and calculating the average distance of the sentence vectors of each sentence set, the average distance of the sentence vectors of the sentence set is the average value of the distance between every two sentence vectors in the sentence set; calculating the mean of the average distance of the sentence vectors of each sentence set to obtain the distance threshold.

[0016] According to another aspect of an embodiment of the present invention, a device for determining a business problem is provided.

[0017] A device for determining business problems includes: a semantic analysis module, which is used to perform semantic analysis on acquired audit opinion data to obtain and save sentence vectors of the audit opinion data; a clustering module, which is used to query sentence vectors that meet conditions from the sentence vectors of the stored audit opinion data and cluster the queried sentence vectors; and a business problem determination module, which is used to determine the business problem to be solved based on the sentence vectors of each category obtained by clustering.

[0018] Optionally, the semantic analysis module is also used to: regularly query and obtain the newly added audit opinion data within the most recent preset time period through a scheduled task, perform the semantic analysis on each audit opinion data to obtain the sentence vector of the newly added audit opinion data, and save it to a database, which is used to store the sentence vector of the stored audit opinion data.

[0019] Optionally, the semantic analysis module performs semantic analysis on a review opinion data in the following manner: segmenting the review opinion data to obtain a segmentation array; using a word vector model to calculate a word vector for each word in the segmentation array, and performing weighted averaging on the obtained word vectors according to the weights corresponding to the parts of speech to obtain a sentence vector for the review opinion data.

[0020] Optionally, the business problem determination module is further used to: determine the business problem indicated by the category with the largest number of sentence vectors among the sentence vectors of each category obtained by clustering as the business problem to be solved.

[0021] Optionally, the business problem determination module is also used to: calculate the distance between each sentence vector in each category of sentence vectors obtained by clustering and the center point vector of the category to which it belongs; determine the convergence degree of the sentence vectors of each category based on the distance and the number of sentence vectors of each category; and determine the business problem indicated by the category whose convergence degree meets preset conditions as the business problem to be solved.

[0022] Optionally, the business problem determination module includes a convergence determination submodule, which is used to: sort the sentence vectors of each category according to the distance from the center point vector of the corresponding category to obtain a sentence vector sequence of each category; for each category sentence vector sequence, calculate the proportion of the number of sentence vectors in the category sentence vector sequence whose distance is less than the distance threshold in the total number of sentence vectors of the category to obtain the convergence of the category sentence vector.

[0023] Optionally, it also includes a distance threshold determination module, which is used to determine the distance threshold in the following manner: forming multiple sentence sets, each sentence set includes multiple sentences with similar semantics; converting the sentences in the sentence sets into sentence vectors respectively, and calculating the average distance of the sentence vectors of each sentence set, the average distance of the sentence vectors of the sentence set is the average value of the distance between every two sentence vectors in the sentence set; calculating the mean of the average distance of the sentence vectors of each sentence set to obtain the distance threshold.

[0024] According to yet another aspect of the embodiments of the present invention, an electronic device is provided.

[0025] An electronic device includes: one or more processors; a memory for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method for determining business problems provided by an embodiment of the present invention.

[0026] According to yet another aspect of an embodiment of the present invention, a computer readable medium is provided.

[0027] A computer readable medium stores a computer program, which, when executed by a processor, implements a method for determining a business problem provided by an embodiment of the present invention.

[0028] One embodiment of the above invention has the following advantages or beneficial effects: semantic analysis is performed on the acquired audit opinion data to obtain and save the sentence vectors of the audit opinion data; the sentence vectors that meet the conditions are queried from the sentence vectors of the stored audit opinion data, and the queried sentence vectors are clustered; the business problem to be solved is determined based on the sentence vectors of each category obtained by clustering. The ability to automatically identify existing potential problem points can effectively reduce cognitive biases caused by subjective reasons when manually analyzing business problems, directly hit business pain points, avoid waste of R&D resources, and reduce the number of information transmissions between people, thereby reducing information loss, improving the accuracy of positioning needs, and discovering potential problems before the problem actually occurs to optimize the system, avoiding business problems from being displayed to users and affecting user experience.

[0029] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0031] Figure 1 is a schematic diagram of main steps of a method for determining a business problem according to an embodiment of the present invention;

[0032] Figure 2is a schematic diagram of a process for determining a business problem according to an embodiment of the present invention;

[0033] Figure 3 is a schematic diagram of main modules of an apparatus for determining a business problem according to an embodiment of the present invention;

[0034] Figure 4 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0035] Figure 5 It is a schematic diagram of the structure of a computer system suitable for implementing a server of an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0037] Figure 1 It is a schematic diagram of main steps of a method for determining a business problem according to an embodiment of the present invention.

[0038] like Figure 1 As shown, a method for determining a business problem according to an embodiment of the present invention mainly includes the following steps S101 to S103.

[0039] Step S101: perform semantic analysis on the acquired audit opinion data to obtain and save sentence vectors of the audit opinion data.

[0040] Step S102: querying sentence vectors that meet the conditions from the sentence vectors of the stored audit opinion data, and clustering the queried sentence vectors.

[0041] Step S103: Determine the business problem to be solved based on the sentence vectors of each category obtained by clustering.

[0042] There will be various audit functions in the operation system, and each audit record will inevitably contain a very important information called "audit opinion", which can reflect the irregularities in customer operations, and can also indirectly reflect the unreasonable, misleading, and problematic process design points in the application. The audit opinion data of the embodiment of the present invention is the "audit opinion", for example: user real-name authentication information audit, user-generated large transaction order audit, user registration submission of enterprise qualifications audit, user application for the opening of a certain authority audit, user application for invoice audit, etc. The audit opinion data of the embodiment of the present invention is not limited to the audit opinions listed above, and can be audit opinions in various scenarios that require audit. It is generated by the operation team in daily operation audit work, and is not proposed from the user side. The embodiment of the present invention finds areas that can be optimized from the operation audit opinions to actively optimize the application, thereby avoiding the defects caused by the process of waiting for users to feedback problems to the operation team and then the operation team feedbacks problems to the R&D team and then optimizes.

[0043] In one embodiment, a scheduled task can be used to regularly query and obtain the newly added audit opinion data within the most recent preset time period, and semantic analysis can be performed on each audit opinion data to obtain the sentence vector of the newly added audit opinion data, and save it to a database, which is used to store the sentence vector of the stored audit opinion data. The sentence vector of the stored audit opinion data refers to the sentence vector of the audit opinion data stored in the database.

[0044] Semantic analysis of an audit opinion data can be performed in the following way: the audit opinion data is segmented to obtain a segmentation array; a word vector model is used to calculate a word vector for each word in the segmentation array, and the obtained word vectors are weighted averaged according to the weights corresponding to the parts of speech to obtain a sentence vector for the audit opinion data.

[0045] The word segmentation operation can be implemented by various word segmentation algorithms, and the word vector model includes but is not limited to the word2vec model. Each word obtained by word segmentation has a corresponding part of speech, and each part of speech has a corresponding weight. The size of the weight is related to the part of speech. Nouns and verbs can express the main semantics of the sentence, so they have higher weights, while words of other parts of speech have lower weights. Specifically, the weight of each part of speech can be set according to the above weight setting principles as needed.

[0046] In one embodiment, determining the business problem to be solved based on the sentence vectors of each category obtained by clustering includes: determining the business problem indicated by the category with the largest number of sentence vectors among the sentence vectors of each category obtained by clustering as the business problem to be solved.

[0047] In another embodiment, a business problem to be solved is determined based on sentence vectors of each category obtained by clustering, including: respectively calculating the distance between each sentence vector of each category obtained by clustering and the center point vector of its category; determining the convergence of the sentence vectors of each category based on the distance (i.e., the distance between each sentence vector of each category and the center point vector of its category) and the number of sentence vectors of each category; and determining the business problem indicated by the category whose convergence meets a preset condition as the business problem to be solved.

[0048] Determining the convergence of sentence vectors of each category according to the distance between each sentence vector in each category and the center point vector of its category and the number of sentence vectors of each category, specifically may include: sorting the sentence vectors of each category according to the distance from the center point vector of the corresponding category to obtain a sequence of sentence vectors of each category; for each category sentence vector sequence, respectively calculating the proportion of the number of sentence vectors in the category sentence vector sequence whose distance is less than the distance threshold in the total number of sentence vectors of the category, to obtain the convergence of the sentence vector of the category.

[0049] The distance threshold can be determined in the following manner: forming multiple sentence sets, each sentence set includes multiple sentences with similar semantics; converting the sentences in the sentence set into sentence vectors respectively, and calculating the average distance of the sentence vectors of each sentence set, the average distance of the sentence vectors of the sentence set is the average of the distances between every two sentence vectors in the sentence set; calculating the mean of the average distances of the sentence vectors of each sentence set to obtain the distance threshold. Preferably, the semantics of different sentence sets should be different, that is, each sentence set corresponds to a topic, and the topics of different sentence sets are different.

[0050] The preset conditions used to determine the business problem to be solved can be set as needed, for example, setting the business problem indicated by the category with a convergence degree greater than or equal to a certain value (between 0 and 1, such as 0.8) as the business problem to be solved.

[0051] Figure 2 The figure is a schematic diagram of a process for determining a business problem according to an embodiment of the present invention.

[0052] like Figure 2 As shown, the process of determining business problems in one embodiment of the present invention can be executed on the server side, and can be specifically implemented using a data analysis server and an operation system (related servers of the operation system). The data analysis server and the operation system server can be the same server or different servers.

[0053] Specifically, the audit opinion data is obtained through the operation system and semantic analysis is performed to obtain the sentence vectors of the audit opinion data, which are then saved in the database. The data analysis server includes a data analysis tool (hereinafter referred to as the analysis tool), which is used to query the sentence vectors that meet the conditions from the database according to the specified conditions, and cluster the queried sentence vectors, calculate the distance between each audit opinion and the center point of its class, that is, the distance between the sentence vector of the audit opinion and the center point vector of its category, and sort the sentence vectors of each audit opinion in each category from small to large according to the distance to determine the business problem to be solved. The process of this embodiment will be described in detail below.

[0054] Since the model loading and sentence vector calculation of semantic recognition are relatively time-consuming and CPU-consuming processes, and the analysis of audit opinion data does not require high data real-time performance, in order not to affect normal online business, the above processing can be performed in the form of scheduled tasks (for example, the audit opinion data within the previous 24 hours are processed uniformly at 2 a.m. every day). The operating system of this embodiment uses a scheduled task framework to perform semantic analysis on the newly added audit opinion data within a time period, generate sentence vectors for the audit opinion data, and store them in the database. At present, most operating systems are built based on spring. This embodiment can use the stable and mature timer framework quartz to implement it.

[0055] The implementation steps of the scheduled task are as follows: write the job code, which is the content of the specific scheduled task execution, including querying the audit opinion data within the processing cycle, looping through the query results, calculating the sentence vector of the audit opinion data, and storing the sentence vector of the audit opinion data; configure the trigger, which is the trigger of the scheduled task, including the execution time and number of executions of the scheduled task; configure the scheduler, which is the scheduler of the scheduled task and is responsible for associating and binding the job with the trigger; after completing the above steps, directly go online and publish the system, and the scheduled task will begin to be generated.

[0056] The analysis tool will query the audit opinion data that needs to be analyzed according to conditions such as time period, business line, and audit business type, and then analyze it according to the selected clustering algorithm. It will automatically cluster the sentence vectors of the audit opinion data, calculate the distance between the sentence vector of each audit opinion data in the data set composed of the queried audit opinion data and each centroid in the clustering result, and classify them into the category with the closest distance. After calculating the distance between each audit opinion and the center point of its category, the sentence vectors of each audit opinion in each category can be sorted from small to large according to the distance, and finally a statistical chart or statistical analysis report will be generated. Technical product managers can use the statistical chart or statistical analysis report to understand which aspect the audit opinions are concentrated under the specified conditions, so as to put forward system improvement requirements in a more targeted manner.

[0057] Semantic analysis can be implemented based on the word2vec model. Each review opinion data needs to calculate the sentence vector for subsequent clustering calculation. The implementation steps are as follows:

[0058] Use Chinese word segmentation tools to segment the newly added audit opinion data. Currently, there are a variety of open source Chinese word segmenters to choose from. For example, the audit opinion data: "interests and hobbies" will be segmented to get the word segmentation array ["interests", "hobbies"]; each word in the word segmentation array is passed to the model separately and the corresponding word vector is calculated, and then all word vectors are weighted averaged according to the part of speech. Nouns and verbs can represent the main semantics of the sentence, so they can be set to a higher corresponding weight, and other part of speech words can be set to a lower corresponding weight. The value obtained after weighted average is the sentence vector of the audit opinion data. For example:

[0059] Word vector: VEC1 = [v 11 ,v 12 ,...,v 1n ]Weight: K1;

[0060] Word vector: VEC2 = [v 21 ,v 22 ,...,v 2n ]Weight: K2

[0061] …

[0062] Word vector: VEC m =[v m1 ,v m2 ,...,v mn ]Weight: K m

[0063] Sentence vector:

[0064] The calculated sentence vector can be saved in the same database table as the audit opinion data, or a separate database table or even a dedicated distributed database storage can be used. In order not to affect the business processing of the audit opinion data, the second solution is preferably adopted. When storing the sentence vector, the audit opinion data ID (identifier) ​​also needs to be stored at the same time for association.

[0065] The main functions of the data analysis tool are data query, clustering calculation, result analysis, and generating statistical charts or statistical analysis reports. The implementation steps are as follows: call the database interface to query the sentence vector according to the specified conditions to obtain the data set. The query conditions (i.e., specified conditions) can be customized according to the needs of the user, which can be the audit business type, audit time, etc., as long as the database has storage. The query result is the audit opinion data ID and sentence vector. If the data volume of the data set is not large, it can be directly placed in the memory. If the data volume is large, it can be placed in a dedicated cloud storage system; read the data set, perform clustering calculation on the data set, and the clustering algorithm and the number of clusters (number of categories) can be selected by the user. The clustering calculation result is the center point vector of each class; calculate the distance between each sentence vector in the data set and each centroid in the clustering result, and then classify it as the class with the smallest distance, and finally generate a clustering statistics table, in which each category can be understood as a certain type of business problem.

[0066] In one embodiment, product managers can find the business issues with the most audit opinions based on the cluster statistics table, and use this as a breakthrough point for analysis to propose new requirements, thereby making it easier to solve real business pain points.

[0067] In another embodiment, the distance between each sentence vector in each category of sentence vectors obtained by clustering and the center point vector of its category can be calculated respectively, and then the sentence vectors of each category are sorted according to the distance with the center point vector of the corresponding category to obtain a sentence vector sequence of each category, and for each category of sentence vector sequence, the proportion of the number of sentence vectors in the category sentence vector sequence whose distance from the center point vector of the category is less than a distance threshold in the total number of sentence vectors of the category is calculated respectively as the convergence degree of the category sentence vector, and it is judged whether the convergence degree of each category sentence vector meets the preset condition. If the convergence degree of a category sentence vector meets the preset condition, then the business problem indicated by the category can be determined as the business problem to be solved.

[0068] Using the convergence of sentence vectors to determine the business problems to be solved can improve the accuracy of the results. For example, there are a total of 100 audit opinion data. After clustering calculation, the three types of business problems A, B, and C correspond to 50, 30, and 20 respectively. If the sentence vectors of the 50 audit opinion data in A are far away from the center point vector of A and are relatively discrete, while the sentence vectors of the 30 audit opinion data in B are relatively close to the center point vector of B and are relatively convergent, then although the sentence vectors of the audit opinion data in category A are large in number, they may be audit opinion data with a large semantic gap from categories A, B, and C, but are classified into category A just because they are closer to the center point of category A. In fact, the clustering results in category B are more convergent and closer to the actual semantics. Therefore, the system should be optimized for the business problems indicated by category B first. By calculating the convergence of sentence vectors of each category, the business problems to be solved can be more accurately determined based on the convergence.

[0069] The clustering algorithm of the embodiment of the present invention can be various clustering algorithms, and the k-means algorithm is introduced below. The clustering operation implementation steps of an embodiment of the present invention are as follows: first, it is necessary to select a k value, that is, the number of clustering categories. The selection of the k value has a great influence on the result. There are usually two selection methods: one is the elbow method, in short, judging the best effect when k is determined based on the functional relationship between the clustering result and k; the other is determined according to specific needs. For example, when clustering a certain product, it may be considered that the product needs to be divided into several categories according to a certain attribute in order to set the k value; select the initial cluster point (or centroid), which is usually randomly selected within the data range. There are two processing methods: one is to take the average multiple times, and the other is bisecting K-means. The main idea of ​​the bisection k-means algorithm is: first, all points are taken as a cluster, and then the cluster is divided into two. Then, the cluster that can minimize the clustering cost function (i.e., the sum of squared errors) is selected and divided into two clusters, and this process is continued until the number of clusters is equal to the k value; then, the distances from these centroids are calculated for all points in the data set, and they are divided into the category closest to their centroids. After completion, the average value of each cluster is calculated, and this point is used as the new centroid. The two steps are repeated until the new centroid is exactly the same as the original centroid to obtain the final result. The centroid that is finally stabilized after repeated calculations by the clustering algorithm is the center point vector.

[0070] The process of determining business problems in the embodiment of the present invention analyzes the audit opinion data in the operation system through a combination of a series of artificial intelligence algorithms, so as to find out the points where product problems are concentrated. On this basis, the technical product manager directly generates a PRD (product requirement document), which effectively reduces the cognitive bias caused by subjective reasons when the product manager of the operation team and the product manager of the technical team analyze business problems, reduces the information loss caused by transmission, and enables PRD to directly hit the business pain points and avoid waste of R&D resources. At the same time, it can discover potential problems before the problems actually occur, optimize the system, nip business problems in the bud, and effectively improve user experience.

[0071] Figure 3 4 is a schematic diagram of main modules of an apparatus for determining a business problem according to an embodiment of the present invention.

[0072] like Figure 3 As shown, an apparatus 300 for determining a business problem in one embodiment of the present invention mainly includes: a semantic analysis module 301 , a clustering module 302 , and a business problem determination module 303 .

[0073] The semantic analysis module 301 is used to perform semantic analysis on the acquired audit opinion data to obtain and save the sentence vectors of the audit opinion data.

[0074] The clustering module 302 is used to query the sentence vectors that meet the conditions from the sentence vectors of the stored review opinion data, and cluster the queried sentence vectors.

[0075] The business problem determination module 303 is used to determine the business problem to be solved according to the sentence vectors of each category obtained by clustering.

[0076] There will be various audit functions in the operation system. Each audit record must contain a very important information called "audit opinion", which can reflect the irregularities in customer operations, and can also indirectly reflect the unreasonable, misleading, and problematic points in the application process design. The audit opinion data in the embodiment of the present invention is the "audit opinion".

[0077] In one embodiment, the semantic analysis module 301 periodically queries and obtains the newly added audit opinion data within the most recent preset time period through a scheduled task, performs the semantic analysis on each audit opinion data to obtain the sentence vector of the newly added audit opinion data, and saves it to a database, which is used to store the sentence vectors of the above-mentioned stored audit opinion data.

[0078] The semantic analysis module 301 can perform semantic analysis on a review opinion data in the following manner: segment the review opinion data into words to obtain a segmentation array; use the word vector model to calculate the word vector for each word in the segmentation array, and perform weighted averaging on the obtained word vectors according to the weights corresponding to the parts of speech to obtain the sentence vector of the review opinion data.

[0079] In one embodiment, the business problem determination module 303 may be specifically configured to: determine the business problem indicated by the category with the largest number of sentence vectors among the sentence vectors of each category obtained by clustering as the business problem to be solved.

[0080] In another embodiment, the business problem determination module 303 can be specifically used to: respectively calculate the distance between each sentence vector in each category of sentence vectors obtained by clustering and the center point vector of its category; determine the convergence degree of each category of sentence vectors based on the distance and the number of sentence vectors in each category; and determine the business problem indicated by the category whose convergence degree meets the preset conditions as the business problem to be solved.

[0081] Among them, the business problem determination module 303 may include a convergence determination submodule, which is used to: sort the sentence vectors of each category according to the distance from the center point vector of the corresponding category to obtain a sentence vector sequence of each category; for each category sentence vector sequence, calculate the proportion of the number of sentence vectors in the category sentence vector sequence whose distance from the center point of the category is less than the distance threshold in the total number of sentence vectors of the category to obtain the convergence of the category sentence vector.

[0082] The device 300 for determining business problems may also include a distance threshold determination module, which is used to determine the distance threshold in the following manner: forming multiple sentence sets, each sentence set including multiple sentences with similar semantics; converting the sentences in the sentence sets into sentence vectors respectively, and calculating the average distance of the sentence vectors of each sentence set, the average distance of the sentence vectors of the sentence set is the average value of the distance between every two sentence vectors in the sentence set; calculating the mean of the average distance of the sentence vectors of each sentence set to obtain the distance threshold.

[0083] The embodiment of the present invention uses a combination of a series of artificial intelligence algorithms to regularly analyze the audit opinion data in the operation system, so as to find out the potential problem points of the product, which can effectively reduce the cognitive bias caused by subjective reasons when analyzing business problems, so that the product requirement document can directly hit the business pain points, avoid the waste of R&D resources, reduce the number of times information is transmitted between people, thereby reducing information loss, and improving the accuracy of requirements. It can discover potential problems before the problems actually occur, optimize the system, avoid exposing business problems to users, and improve user experience.

[0084] In addition, the specific implementation content of the device for determining a business problem in the embodiment of the present invention has been described in detail in the method for determining a business problem described above, so the repeated content will not be described again here.

[0085] Figure 4 An exemplary system architecture 400 is shown to which the method for determining a business problem or the apparatus for determining a business problem according to an embodiment of the present invention can be applied.

[0086] like Figure 4 As shown, system architecture 400 may include terminal devices 401, 402, 403, network 404 and server 405. Network 404 is used to provide a medium for communication links between terminal devices 401, 402, 403 and server 405. Network 404 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0087] Users can use terminal devices 401, 402, 403 to interact with server 405 through network 404 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 401, 402, 403, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only as examples).

[0088] The terminal devices 401 , 402 , and 403 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0089] The server 405 may be a server that provides various services, such as a backend management server (only an example) that provides support for shopping websites browsed by users using the terminal devices 401, 402, and 403. The backend management server may analyze and process the received data such as product information query requests, and feed back the processing results (such as product information - only an example) to the terminal device.

[0090] It should be noted that the method for determining a business problem provided in the embodiment of the present invention is generally executed by the server 405 , and accordingly, the device for determining a business problem is generally disposed in the server 405 .

[0091] It should be understood that Figure 4 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0092] Reference below Figure 5 , which shows a schematic diagram of the structure of a computer system 500 suitable for implementing a server of an embodiment of the present application. Figure 5The server shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0093] like Figure 5 As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage part 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the system 500 are also stored. The CPU 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0094] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read therefrom is installed into the storage section 508 as needed.

[0095] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the central processing unit (CPU) 501, the above-mentioned functions defined in the system of the present application are executed.

[0096] It should be noted that the computer-readable medium shown in the present invention can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0097] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0098] The modules involved in the embodiments of the present invention may be implemented in software or hardware. The modules described may also be provided in a processor. For example, they may be described as follows: a processor includes a semantic analysis module, a clustering module, and a business problem determination module. The names of these modules do not, in some cases, constitute limitations on the modules themselves. For example, the semantic analysis module may also be described as a "module for performing semantic analysis on the acquired audit opinion data to obtain and save the sentence vectors of the audit opinion data."

[0099] As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently without being assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device includes: performing semantic analysis on the acquired audit opinion data to obtain and save the sentence vectors of the audit opinion data; querying the sentence vectors that meet the conditions from the sentence vectors of the stored audit opinion data, and clustering the queried sentence vectors; determining the business problem to be solved based on the sentence vectors of each category obtained by clustering.

[0100] According to the technical solution of the embodiment of the present invention, semantic analysis is performed on the acquired audit opinion data to obtain and save the sentence vectors of the audit opinion data; the sentence vectors that meet the conditions are queried from the sentence vectors of the stored audit opinion data, and the queried sentence vectors are clustered; the business problems to be solved are determined based on the sentence vectors of each category obtained by clustering. The ability to automatically identify existing potential problem points can effectively reduce cognitive biases caused by subjective reasons when manually analyzing business problems, directly hit business pain points, avoid waste of R&D resources, and reduce the number of information transmissions between people, thereby reducing information loss and improving the accuracy of requirements. Potential problems can be discovered before they actually occur to optimize the system and avoid business problems being displayed to users and affecting user experience.

[0101] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for determining a business problem, characterized in that include: Perform semantic analysis on the obtained audit opinion data to obtain and save the sentence vector of the audit opinion data; Query the sentence vectors that meet the conditions from the sentence vectors of the stored audit opinion data, and cluster the queried sentence vectors; Determine the business problem to be solved based on the sentence vectors of each category obtained by clustering; The determining the business problem to be solved according to the sentence vectors of each category obtained by clustering includes: determining the business problem indicated by the category with the largest number of sentence vectors among the sentence vectors of each category obtained by clustering as the business problem to be solved; Alternatively, the distance between each sentence vector in each category of sentence vectors obtained by clustering and the center point vector of the category to which it belongs is calculated respectively, and the convergence degree of the sentence vectors of each category is determined according to the distance and the number of sentence vectors of each category, and the business problem indicated by the category whose convergence degree meets the preset conditions is determined as the business problem to be solved.

2. The method according to claim 1, characterized in that The newly added audit opinion data within the most recent preset time period is regularly queried and acquired through a scheduled task, and the semantic analysis is performed on each audit opinion data to obtain the sentence vector of the newly added audit opinion data, and the sentence vector is saved in a database, and the database is used to store the sentence vector of the stored audit opinion data.

3. The method according to claim 1 or 2, characterized in that: The semantic analysis of the audit opinion data is performed in the following way: Perform word segmentation on the audit opinion data to obtain a word segmentation array; The word vector model is used to calculate the word vector for each word in the word segmentation array, and the obtained word vectors are weighted averaged according to the weights corresponding to the parts of speech to obtain the sentence vector of the review opinion data.

4. The method according to claim 1, characterized in that: The determining, according to the distance and the number of the sentence vectors of each category, the convergence degree of the sentence vectors of each category includes: Sort the sentence vectors of each category according to the distance from the center point vector of the corresponding category to obtain the sentence vector sequence of each category; For each category sentence vector sequence, the proportion of the number of sentence vectors in the category sentence vector sequence whose distance is less than the distance threshold in the total number of the category sentence vectors is calculated to obtain the convergence degree of the category sentence vector.

5. The method according to claim 4, characterized in that The distance threshold is determined as follows: Form multiple sentence sets, each sentence set includes multiple sentences with similar semantics; Convert the sentences in the sentence set into sentence vectors respectively, and calculate the average distance of the sentence vectors of each sentence set, where the average distance of the sentence vectors of the sentence set is the average value of the distance between every two sentence vectors in the sentence set; The average distance of the sentence vectors of each sentence set is calculated to obtain the distance threshold.

6. A device for determining a business problem, characterized in that: include: A semantic analysis module, used to perform semantic analysis on the obtained audit opinion data to obtain and save the sentence vectors of the audit opinion data; A clustering module is used to query the sentence vectors that meet the conditions from the sentence vectors of the stored review opinion data, and cluster the queried sentence vectors; A business problem determination module is used to determine the business problem to be solved based on the sentence vectors of each category obtained by clustering; The business problem determination module is also used to: determine the business problem indicated by the category with the largest number of sentence vectors among the sentence vectors of each category obtained by clustering as the business problem to be solved; or, respectively calculate the distance between each sentence vector of each category obtained by clustering and the center point vector of its category, determine the convergence degree of the sentence vectors of each category based on the distance and the number of sentence vectors of each category, and determine the business problem indicated by the category whose convergence degree meets the preset conditions as the business problem to be solved.

7. The device according to claim 6, characterized in that The semantic analysis module is also used to: regularly query and obtain the newly added audit opinion data within the most recent preset time period through a scheduled task, perform the semantic analysis on each audit opinion data to obtain the sentence vector of the newly added audit opinion data, and save it to a database, which is used to store the sentence vector of the stored audit opinion data.

8. The device according to claim 6 or 7, characterized in that The semantic analysis module performs semantic analysis on the audit opinion data in the following manner: Perform word segmentation on the audit opinion data to obtain a word segmentation array; The word vector model is used to calculate the word vector for each word in the word segmentation array, and the obtained word vectors are weighted averaged according to the weights corresponding to the parts of speech to obtain the sentence vector of the review opinion data.

9. The device according to claim 6, characterized in that The business problem determination module includes a convergence determination submodule, which is used to: Sort the sentence vectors of each category according to the distance from the center point vector of the corresponding category to obtain the sentence vector sequence of each category; For each category sentence vector sequence, the proportion of the number of sentence vectors in the category sentence vector sequence whose distance is less than the distance threshold in the total number of the category sentence vectors is calculated to obtain the convergence degree of the category sentence vector.

10. The device according to claim 9, characterized in that It also includes a distance threshold determination module, which is used to determine the distance threshold in the following manner: Form multiple sentence sets, each sentence set includes multiple sentences with similar semantics; Convert the sentences in the sentence set into sentence vectors respectively, and calculate the average distance of the sentence vectors of each sentence set, where the average distance of the sentence vectors of the sentence set is the average value of the distance between every two sentence vectors in the sentence set; The average distance of the sentence vectors of each sentence set is calculated to obtain the distance threshold.

11. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 5.

12. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images