Construction method of user experience knowledge base and construction device of user experience knowledge base

Abstract

The invention provides a construction method of a user experience knowledge base and a construction device of the user experience knowledge base. The method comprises the steps that user experience data are obtained, the user experience data are reaction data generated in the interaction process of a user and a product, and the user experience data comprise user feeling data, product performance data, product feature data and use scene data; according to the use scene data, classifying the multiple pieces of user experience data to obtain multiple data sets; according to the multiple data sets, a target user experience network is constructed, a user experience knowledge base is obtained, and the target user experience network is used for representing the influence probability of any one of the product performance data, the product feature data and the use scene data on the user experience data; and obtaining a target search condition, and searching in the user experience knowledge base according to the target search condition to obtain a target search result. According to the method, the problem that user experience analysis is subjective in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of user experience, and in particular, to a method for constructing a user experience knowledge base, a device for constructing a user experience knowledge base, and a computer-readable storage medium. Background Art

[0002] User Experience (UX) can help product and service designers discover important customer experience factors, such as user preferences, usage environment, product features, service performance and their relationships. In the prior art, traditional methods are generally used to collect customer experience information, such as questionnaires, user experience evaluations, and experience tasks. However, these data are usually limited by the initial question settings, and they do not make it easy for designers to identify all key experience factors, such as user preferences, backgrounds, related product features, etc. In addition, the user experience analysis process is highly subjective.

[0003] Therefore, there is an urgent need for a method to solve the problem that user experience analysis in the prior art is relatively subjective. Summary of the invention

[0004] The main purpose of the present application is to provide a method for constructing a user experience knowledge base, a device for constructing a user experience knowledge base, and a computer-readable storage medium, so as to at least solve the problem that user experience analysis in the prior art is relatively subjective.

[0005] According to one aspect of the present application, a method for constructing a user experience knowledge base is provided, comprising: obtaining user experience data, wherein the user experience data is reaction data generated during the interaction between a user and a product, and the user experience data includes user feeling data, product performance data, product feature data, and usage scenario data, wherein the user feeling data is used to characterize the subjective feelings of the user in the process of using the product, the product performance data is used to characterize the performance indicators of the product in the process of the user using the product, the product feature data is used to characterize the functional features of the product in the process of the user using the product, and the usage scenario data is used to characterize the corresponding scenarios in the process of the user using the product; and performing analysis on a plurality of user experience data according to the usage scenario data. Classify to obtain multiple data sets, wherein any one of the data sets includes multiple data groups, and any one of the data groups includes multiple user experience data; construct a target user experience network based on the multiple data sets to obtain a user experience knowledge base, wherein the target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data; obtain a target search condition, search in the user experience knowledge base based on the target search condition, and obtain a target search result, wherein the target search condition is the user experience data to be processed, and the target search result includes the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data to be processed.

[0006] Optionally, obtaining user experience data includes: obtaining user review data and product usage data, wherein the user review data includes the user's review text on the product, and the product usage data includes the process data generated by the user in the process of using the product; determining the user experience data based on the user review data; performing word segmentation processing on the review text to obtain multiple initial words and initial phrases, wherein the review text includes multiple initial sentences, and each of the initial sentences includes multiple initial words; performing part-of-speech tagging processing on the multiple initial words respectively to obtain the part-of-speech corresponding to each of the initial words, and performing part-of-speech tagging processing on the multiple initial phrases respectively to obtain the part-of-speech corresponding to each of the initial phrases, wherein the part-of-speech includes at least noun, verb, adjective and preposition; determining that the initial word corresponding to the verb and the initial word corresponding to the adjective are first preliminary opinion words, and determining the positive words whose number is in the top x among multiple positive words The first preliminary opinion words, the synonyms of the positive words whose number is in the top x, the negative words whose number is in the top x among the multiple negative words, and the synonyms of the negative words whose number is in the top x are the second preliminary opinion words, wherein x is a positive integer less than a predetermined value, and the predetermined value is the minimum value of the number of the positive words and the number of the negative words; the same initial words in the first preliminary opinion words and the second preliminary opinion words are determined as target opinion words, and the initial sentences corresponding to the target opinion words are determined as target opinion sentences; the initial words and the initial phrases corresponding to the noun whose part of speech is determined as first candidate words, and the initial phrases corresponding to the preposition, the verb, and the noun whose part of speech is determined as second candidate words; the product feature data is determined at least based on the target opinion sentence and the first candidate words; the usage scenario data is determined at least based on the product feature data and the second candidate words; the product performance data is determined based on the product usage data.

[0007] Optionally, determining the user feeling data according to the user comment data includes: obtaining historical user comment data and user feeling data corresponding to the historical user comment data, wherein the historical user comment data is the user comment data obtained before the current moment, and the user feeling data includes the emotional tendency of the user; inputting a plurality of the user comment data into a target neural network in sequence to obtain a plurality of user feeling data corresponding to the user comment data, wherein the target neural network is trained by machine learning using a plurality of sets of historical data, and each set of the plurality of historical data includes: the historical user comment data and user feeling data corresponding to the historical user comment data.

[0008] Optionally, determining the product feature data at least based on the target opinion sentence and the first candidate word includes: calculating the word frequency of the first candidate word in the user review data to obtain a first word frequency; obtaining the number of the target opinion sentences that do not include the first candidate word in a plurality of the target opinion sentences to obtain a first number; determining that the initial sentence that is not the target opinion sentence in a plurality of the initial sentences is a non-target opinion sentence; obtaining the number of the non-target opinion sentences that include the first candidate word in a plurality of the non-target opinion sentences to obtain a second number; calculating the distance between the initial sentence containing the first candidate word and the target opinion sentence to obtain a sentence distance; and calculating the distance between the initial sentence containing the first candidate word and the target opinion sentence according to the formula Calculate the weight w of the first candidate word, where t is the first word frequency, n1 is the first number, n2 is the second number, and d is the sentence distance; sort the weights of multiple first candidate words, and obtain y first candidate words with the largest weights as the product feature data.

[0009] Optionally, the usage scenario data is determined based at least on the product feature data and the second candidate words, including: determining the initial sentence including the target opinion word and at least one of the first candidate words as a candidate sentence, and determining the candidate sentence including the second candidate word as a target sentence; obtaining any two of the multiple second candidate words in the target sentence to obtain a candidate word group; calculating the conditional symmetric probability of the two second candidate words in each candidate word group to obtain multiple conditional symmetric probabilities; and determining the second candidate word corresponding to the maximum value of the multiple conditional symmetric probabilities as the usage scenario data.

[0010] Optionally, the product performance data is determined based on the product usage data, including: dividing the process data to obtain multiple process nodes; obtaining the contact points corresponding to each of the process nodes to obtain multiple contact points, wherein the contact points are used to characterize the interaction process between the product and the user; obtaining influencing factor data of each of the contact points to obtain the product performance data, wherein the influencing factor data is used to characterize data that affects the user experience of the user.

[0011] Optionally, according to the usage scenario data, a plurality of the user experience data are classified to obtain a plurality of data sets, including: according to the formula Calculate the local connection M l , where |s j ∩s k | is the jth usage scenario data s j and the kth usage scenario data s k The number of identical language fragments in |sj ∪s k | is the jth usage scenario data s j The number of language segments and the k-th usage scenario data s k The sum of the number of language segments, including words, phrases and sentences; according to formula M g + = sim(r aj +r bk ), calculate the global connection M g , where r aj To include j The a-th user comment data, r bk To include k The b-th user comment data, sim(r aj +r bk ) is r aj and r bk Similarity; According to the formula M = α × M l +β×M g , a connection matrix M is calculated, where α, β∈[0,1], α+β=1, and the connection matrix is ​​used to characterize the association relationship between the multiple data sets.

[0012] Optionally, constructing a user experience network based on the plurality of data sets to obtain a user experience knowledge base includes: constructing an initial directed acyclic graph based on the data sets, wherein the directed acyclic graph includes four data node groups, each of the data node groups includes a plurality of data nodes, and the data node groups correspond one to one with one of the user experience data, product performance data, product feature data, and usage scenario data; determining whether any two of the user experience data, product feature data, and usage scenario data are located in the same initial statement, and in the case where any two of the user experience data, product feature data, and usage scenario data are located in the same initial statement, establishing a connection between two corresponding data nodes, the corresponding data nodes being any two corresponding data nodes of the user experience data, product feature data, and usage scenario data located in the same initial statement; and according to the formula Calculate the value p of the connection to update the initial directed acyclic graph to obtain a preliminary user experience network, where z e is the corresponding data node of the eth node, z f is the corresponding data node of the fth node, |q(z e , z f )| is the number of initial statements in the plurality of initial statements that also include the user experience data corresponding to the corresponding data node, |q(z f )| for multiple initial statements, including zf The number of initial statements of the corresponding user experience data; using a multivariate linear regression analysis method to process the product performance data and the user experience data, and obtain the connection relationship between the data nodes corresponding to the product performance data and the data nodes corresponding to the user experience data, so as to update the preliminary user experience network and obtain the target user experience network.

[0013] According to another aspect of the present application, a device for constructing a user experience knowledge base is provided, comprising: an acquisition unit, for acquiring user experience data, wherein the user experience data is reaction data generated during the interaction between a user and a product, and the user experience data includes user feeling data, product performance data, product feature data, and usage scenario data, the user feeling data is used to characterize the subjective feelings of the user in the process of using the product, the product performance data is used to characterize the performance indicators of the product in the process of the user using the product, the product feature data is used to characterize the functional features of the product in the process of the user using the product, and the usage scenario data is used to characterize the corresponding scenarios in the process of the user using the product; a classification unit, for classifying a plurality of user experience data according to the usage scenario data. The invention relates to a method for performing row classification to obtain a plurality of data sets, wherein any one of the data sets includes a plurality of data groups, and any one of the data groups includes a plurality of the user experience data; a construction unit is used to construct a target user experience network based on the plurality of data sets to obtain a user experience knowledge base, wherein the target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data; a search unit is used to obtain a target search condition, and search the user experience knowledge base according to the target search condition to obtain a target search result, wherein the target search condition is the user experience data to be processed, and the target search result includes the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data to be processed.

[0014] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the methods described.

[0015] Applying the technical solution of this application, first, obtain user experience data, which is the reaction data generated during the interaction between users and products. User experience data includes user experience data, product performance data, product feature data, and usage scenario data; then, classify multiple user experience data according to the usage scenario data to obtain multiple data sets; then, construct a target user experience network according to the multiple data sets to obtain a user experience knowledge base, which is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data; finally, obtain the target search condition, search in the user experience knowledge base according to the target search condition, and obtain the target search result. This solution reflects the key factors of user experience data by combining the four aspects of user experience data, product performance data, product feature data, and usage scenario data, so as to expand the user experience data and reduce the limitations; in addition, by classifying the user experience data and constructing the target user experience network, the user experience knowledge base is constructed, and then by searching the user experience knowledge base, the subjectivity of the causal dependency analysis between user experience data can be reduced, solving the problem that the user experience analysis in the prior art is relatively subjective. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings constituting part of the present application are used to provide a further understanding of the present application. The exemplary embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 A hardware structure block diagram of a mobile terminal for executing a method for constructing a user experience knowledge base provided in an embodiment of the present application is shown;

[0018] Figure 2 A schematic diagram of a process for constructing a user experience knowledge base provided according to an embodiment of the present application is shown;

[0019] Figure 3 A schematic diagram of a process for obtaining user experience data provided in an embodiment of the present application is shown;

[0020] Figure 4 A schematic diagram of the structure of a target user experience network provided according to an embodiment of the present application is shown;

[0021] Figure 5 A specific flow chart of a method for constructing a user experience knowledge base provided in accordance with an embodiment of the present application is shown;

[0022] Figure 6 A structural block diagram of a device for constructing a user experience knowledge base provided according to an embodiment of the present application is shown.

[0023] The above drawings include the following reference numerals:

[0024] 102, processor; 104, memory; 106, transmission device; 108, input and output devices. DETAILED DESCRIPTION

[0025] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0026] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] As introduced in the background technology, user experience analysis in the prior art is relatively subjective. To solve the above problem, the embodiments of the present application provide a method for constructing a user experience knowledge base, an apparatus for constructing a user experience knowledge base, and a computer-readable storage medium.

[0029] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0030] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 FIG. 1 is a hardware structure block diagram of a mobile terminal of a method for constructing a user experience knowledge base according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is for illustration only and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0031] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for constructing the user experience knowledge base in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, the above method is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The transmission device 106 is used to receive or send data via a network. The above-mentioned specific examples of the network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0032] In this embodiment, a method for constructing a user experience knowledge base running on a mobile terminal, a computer terminal or a similar computing device is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0033] Figure 2 FIG. 1 is a flow chart of a method for constructing a user experience knowledge base according to an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:

[0034] Step S201, obtaining user experience data, wherein the user experience data is reaction data generated during the interaction between the user and the product, and the user experience data includes user experience data, product performance data, product feature data, and usage scenario data, wherein the user experience data is used to characterize the subjective experience of the user in the process of using the product, the product performance data is used to characterize the performance index of the product in the process of the user using the product, the product feature data is used to characterize the functional features of the product in the process of the user using the product, and the usage scenario data is used to characterize the corresponding scenario in the process of the user using the product;

[0035] Specifically, user experience data can describe the corresponding psychological state of the product in a specific situation, such as positive, neutral and negative opinions. It can be expressed in emotional phrases, such as: easy to use, interesting. Product performance data can describe the key performance indicators of product services to reflect the actual operation of product services. These indicators indirectly affect customer satisfaction, such as: the proportion of automatic approval for credit card opening, the time from loan issuance to account receipt, the one-time pass rate of face recognition, etc. Product feature data refers to the product characteristics involved in user experience. It includes product entities, such as the credit card limit, rights and interests, and the appearance of the card. Usage scenario data can describe the scenario factors when using the product, such as location, time and activity.

[0036] Step S202, classifying the plurality of user experience data according to the usage scenario data to obtain a plurality of data sets, wherein any one of the data sets includes a plurality of data groups, and any one of the data groups includes a plurality of user experience data;

[0037] Specifically, the user experience data integration process will be grouped according to different user experience aspects. In this solution, the focus is on grouping user experience data according to scenarios.

[0038] Step S203, constructing a target user experience network based on the multiple data sets to obtain a user experience knowledge base, wherein the target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data;

[0039] Specifically, the target user experience network is expressed as: C is the product feature data, P is the product performance data, Q is the user experience data, and T is the usage scenario data.

[0040] Step S204, obtaining target search conditions, searching in the user experience knowledge base according to the target search conditions, and obtaining target search results, wherein the target search conditions are user experience data to be processed, and the target search results include the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data to be processed.

[0041] Specifically, after the user experience knowledge base is built, it can be used to further optimize the customer experience. When a large number of user experience questionnaires are collected, it is impossible to analyze them one by one manually. The user experience knowledge base can be used to filter out questionnaires with negative emotions, and then experience optimization personnel can intervene to solve user problems. In addition, product business technicians will re-optimize products or business processes based on the user experience knowledge base. After the optimization, they will observe the user experience knowledge base again after a period of time to determine whether the user experience has improved.

[0042] Through this embodiment, a method for constructing a user experience knowledge base is provided. First, user experience data is obtained. User experience data is the reaction data generated during the interaction between users and products. User experience data includes user experience data, product performance data, product feature data, and usage scenario data. Then, multiple user experience data are classified according to the usage scenario data to obtain multiple data sets. Then, a target user experience network is constructed according to the multiple data sets to obtain a user experience knowledge base. The target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data. Finally, a target search condition is obtained, and a search is performed in the user experience knowledge base according to the target search condition to obtain a target search result. This solution reflects the key factors of user experience data by combining the four aspects of user experience data, product performance data, product feature data, and usage scenario data, so as to expand the user experience data and reduce limitations. In addition, the construction of the user experience knowledge base is realized by classifying the user experience data and constructing the target user experience network. Then, through the search of the user experience knowledge base, the subjectivity of the causal dependency analysis between user experience data can be reduced, solving the problem that the user experience analysis in the prior art is relatively subjective.

[0043] In the specific implementation process, the above step S201 can be implemented by the following steps: step S2011, obtaining user comment data and product usage data, wherein the above user comment data includes the comment text of the above user on the above product, and the above product usage data includes the process data generated by the above user in the process of using the above product; step S2012, determining the above user experience data according to the above user comment data; step S2013, performing word segmentation processing on the above comment text to obtain multiple initial words and initial phrases, wherein the above comment text includes multiple initial sentences, and each of the above initial sentences includes multiple initial words; step S2014, performing part-of-speech tagging processing on the multiple initial words respectively to obtain the part-of-speech corresponding to each of the above initial words, and performing part-of-speech tagging processing on the multiple initial phrases respectively to obtain the part-of-speech corresponding to each of the above initial phrases, wherein the part-of-speech at least includes noun, verb, adjective and preposition; step S2015, determining that the above initial words corresponding to the verb and the above initial words corresponding to the adjective are first preliminary opinion words, and determining that the number of positive words is in the top x The positive words in the first x positions, the synonyms of the positive words in the first x positions, the negative words in the first x positions among the multiple negative words, and the synonyms of the negative words in the first x positions are the second preliminary opinion words, wherein x is a positive integer less than a predetermined value, and the predetermined value is the minimum value of the number of the positive words and the number of the negative words; step S2016, determining the same initial words in the first preliminary opinion words and the second preliminary opinion words as target opinion words, and determining the initial sentences corresponding to the target opinion words as target opinion sentences; step S2017, determining the initial words and initial phrases corresponding to the nouns in the above-mentioned part of speech as first candidate words, and determining the initial phrases corresponding to the prepositions, verbs and nouns in the above-mentioned part of speech as second candidate words; step S2018, determining the product feature data at least based on the target opinion sentences and the first candidate words; step S2019, determining the usage scenario data at least based on the product feature data and the second candidate words; step S201110, determining the product performance data based on the product usage data. This method can quickly obtain the above-mentioned user experience data, product performance data, product feature data and usage scenario data, and further improve the comprehensiveness of the above-mentioned user experience data.

[0044] Specifically, opinion words used to express subjective opinions in customer reviews are extracted. Adjectives are mostly used to express opinions, and verbs are usually used to convey customers' emotional orientation. Therefore, first, adjectives and verbs are extracted from the reviews as candidate opinion words using Part-Of-Speech tagging (POS). Then, based on the results of POS, the top n positive and negative words are manually selected from the adjectives and verbs in the reviews as seed word lists, respectively. Using WordNet, synonyms of seed words are added to expand, and finally a positive and negative word set is formed. The final opinion word is extracted by checking whether the candidate opinion word is included in the positive and negative word set, and the comment sentence with any opinion word is regarded as an opinion sentence. Then, given a set of user reviews as input, the user experience discovery process aims to discover terms or phrases that are highly related to the user experience through language unit recognition, namely user experience data, product feature data, and usage scenario data. Secondly, the business operation data is collected from the background system of the product service, and the core data is extracted as the key performance indicator, namely product performance data, to reflect the actual performance of the product or service.

[0045] In order to further quickly obtain the above-mentioned user perception data, the above-mentioned step S2012 of the present application can be implemented through the following steps: step S20121, obtaining historical user comment data and user perception data corresponding to the above-mentioned historical user comment data, wherein the above-mentioned historical user comment data is the above-mentioned user comment data obtained before the current moment, and the above-mentioned user perception data includes the emotional tendency of the above-mentioned user; step S20122, inputting the multiple above-mentioned user comment data into the target neural network in sequence, and obtaining the multiple user perception data corresponding to the above-mentioned user comment data, wherein the above-mentioned target neural network is trained by machine learning using multiple groups of historical data, and each group of the multiple groups of historical data includes: the above-mentioned historical user comment data and user perception data corresponding to the above-mentioned historical user comment data.

[0046] Specifically, the process of obtaining the user experience data corresponding to the historical user comment data is as follows: first, use the Chinese word vector to convert the word text into a floating point vector, use jieba word segmentation to accurately segment the user comments, and normalize the length of each comment according to the normal distribution to cut off the long and make up the short. After the data preparation is completed, load the customer comment data, 90% of the samples are used for training, and 10% of the samples are used for testing. Build a neural network model for training. The first layer is the word embedding layer, which puts in the word vector matrix. The second layer uses a 64-unit bidirectional long short-term memory network, the third layer uses a 16-unit LSTM, and the last is the full connection layer, using the sigmoid activation function to output the result. Configure the training method, the loss function is the binary cross entropy loss function, the optimization function uses the Adam function, and the model evaluation function uses the accuracy function.

[0047] The above step S2019 can also be implemented in other ways, for example: step S20191, calculate the word frequency of the above first candidate word in the above user comment data to obtain a first word frequency; step S20192, obtain the number of the above target opinion sentences that do not include the above first candidate word in the multiple above target opinion sentences, and obtain a first number; step S20193, determine that the above initial sentences that are not the above target opinion sentences in the multiple above initial sentences are non-target opinion sentences; step S20194, obtain the number of non-target opinion sentences that include the above first candidate word in the multiple above non-target opinion sentences, and obtain a second number; step S20195, calculate the distance between the initial sentence containing the above first candidate word and the above target opinion sentence, and obtain a sentence distance; step S20196, according to the formula Calculate the weight w of the first candidate words, where t is the first word frequency, n1 is the first number, n2 is the second number, and d is the sentence distance; step S20197, sort the weights of the plurality of first candidate words, and obtain the y first candidate words with the largest weights as the product feature data. This method can further quickly determine the product feature data.

[0048] Specifically, if users have more opinions on a particular product feature than others, this means that the feature has a greater impact on the user experience. Therefore, by defining the weight function for calculating the first candidate word as described above, the function assumes that if a noun or noun phrase is closer to any opinion word, that is, the smaller the function value, then it is more likely to become a product feature.

[0049] In some embodiments, the above step S20110 can be specifically implemented by the following steps: step S201101, determining the above initial sentence including the above target opinion word and at least one of the above first candidate words as a candidate sentence, and determining the above candidate sentence including the above second candidate words as a target sentence; step S201102, obtaining any two of the above second candidate words in the above target sentence to obtain a candidate word group; step S201103, calculating the conditional symmetric probabilities of the two second candidate words in each of the above candidate word groups to obtain multiple conditional symmetric probabilities; step S201104, determining the above second candidate word corresponding to the maximum value of the multiple conditional symmetric probabilities as the above usage scenario data. This method can further quickly determine the above usage scenario data.

[0050] Specifically, the selection function measures the phrase pair (v a , v b) to calculate the score of fragment t. The higher the scp, the stronger the correlation. The calculation formula of conditional symmetric probability is as follows: Where n is v a and v b The length, w i Represents v a and v b The i-th word in the corpus, P(w1, w2) represents the probability of a given phrase appearing in the corpus. Then design some checking rules for screening, which use the language characteristics in or between segment groups to select segments. The above checking rules are as follows: check each phrase segment, if the number of words in the phrase segment is less than 2, delete it; group the phrase segments in the set according to whether they have the same phrase pair; in each group, select the phrases with prepositions or v a 、v b The phrase s with the smallest length is added to the set; if the phrase s j Contains phrase segments k , then remove s from the set j .

[0051] Step S201111 can be implemented by the following steps: Step S2011111, dividing the process data to obtain multiple process nodes; Step S2011112, obtaining the contact points corresponding to each of the process nodes to obtain multiple contact points, wherein the contact points are used to characterize the interaction process between the product and the user; Step S2011113, obtaining the influencing factor data of each of the contact points to obtain the product performance data, wherein the influencing factor data is used to characterize the data that affects the user experience of the user. This method can further quickly obtain the product performance data.

[0052] Specifically, by collecting key performance indicators of products or services, the actual operation of the business can be reflected. Confirm the product service process, and split the entire service process into various links according to the service steps. In each link, analyze the contact points between customers and the company through face-to-face, online and other channels. Analyze the factors that may affect the customer experience at each contact point. In response to these influencing factors, collect data indicators that reflect the actual operation and performance of the business. Taking credit card services as an example, the key performance indicators that can be collected are as follows: the length of the credit card opening process, the time from the completion of loan approval to the actual receipt of the loan, the proportion of automatic approval, the one-time pass rate of facial recognition, the equipment failure rate, the customer service manual connection rate, the points redemption rate, etc. The optimization of these key performance indicators can improve customer experience satisfaction. The specific process of obtaining user experience data mentioned above is as follows: Figure 3 shown.

[0053] The above step S202 can be implemented by the following steps: Step S2021, according to the formula Calculate the local connection M l , where |s j ∩s k | is the jth usage scenario data s j and the kth usage scenario data s k The number of identical language fragments in |s j ∪s k | is the jth usage scenario data s j The number of language fragments and the k-th usage scenario data s k The sum of the number of language segments, the above-mentioned language segments include words, phrases and sentences; step S2022, according to formula M g + = sim(r aj +r bk ), calculate the global connection M g , where r aj To include j The a-th user comment data, r bk To include k The b-th user comment data, sim(r aj +r bk ) is r aj and r bk similarity; Step S2023, according to the formula M = α × M l +β×M g , calculate and obtain the connection matrix M, where α, β∈[0,1], α+β=1, and the above connection matrix is ​​used to characterize the association relationship between the above data sets. This method can further quickly realize the classification of user experience data.

[0054] Specifically, the assumption of the above local connection is that if two context features s j and k Sharing more common language fragments indicates that they are more likely to be related to each other. For example, the scenario phrases "when selecting goods" and "when redeeming goods" are closely related to the context of consumers shopping in the points mall. The assumption of global connection is that if two scenario features are often mentioned in similar language environments, then they are likely to refer to similar aspects. For example, when talking about a specific product feature "credit card points", scenario phrases such as "in the mall" and "when redeeming" are usually accompanied by "points". The higher the frequency of occurrence, the higher the possibility that these scenario data are classified as user experience aspect groups related to "credit card points". In practical applications, text can be converted into word vectors, and then the cosine similarity between vectors can be calculated to represent the correlation between scenarios. The k-nearest neighbor algorithm can also be used to group the connection matrix to obtain the corresponding classification.

[0055] The above step S203 can be implemented by the following steps: Step S2031, constructing an initial directed acyclic graph based on the above data set, wherein the above directed acyclic graph includes four data node groups, each of the above data node groups includes multiple data nodes, and the above data node groups correspond one to one with the above user perception data, product performance data, product feature data and usage scenario data; Step S2032, determining whether any two of the above user perception data, product feature data and usage scenario data are located in the same above initial statement, and in the case that any two of the above user perception data, product feature data and usage scenario data are located in the same above initial statement, establishing a connection between two corresponding data nodes, and the above corresponding data nodes are any two corresponding data nodes of the above user perception data, product feature data and usage scenario data located in the same above initial statement; Step S2033, according to the formula Calculate the value p of the above connection to update the above initial directed acyclic graph to obtain a preliminary user experience network, where z e is the eth corresponding data node, z f is the fth corresponding data node, |q(z e , z f )| is the number of initial statements in the above-mentioned initial statements that also include the user experience data corresponding to the above-mentioned corresponding data node, |q(z f )| for multiple of the above initial statements, including z f Step S2034: Process the product performance data and the user experience data using a multivariate linear regression analysis method to obtain the connection relationship between the data nodes corresponding to the product performance data and the data nodes corresponding to the user experience data, so as to update the preliminary user experience network and obtain the target user experience network. This method further quickly constructs the user experience knowledge base by constructing the target user experience network.

[0056] Specifically, Figure 4As shown in the figure, there are four types of nodes in the user experience network, namely, scenario features, i.e., usage scenario data, product features, i.e., product feature data, user emotional state, i.e., user perception data, and key performance indicators, i.e., product performance data. An edge pointing from one node to another node represents the influence of one node on another node. In the three types of nodes, scenario features, product features, and user emotional state, the links between nodes are automatically formed based on text comments. If two nodes appear in the same sentence, an edge is assigned. Based on the influence on user experience, the directions of the edges include: from product features to scenario features; from product features to user emotional state; and from scenario features to user emotional state. In the two types of nodes, i.e., key performance indicators and user emotional state, the links between nodes are obtained through multivariate linear regression analysis, with user emotional state as the dependent variable and key performance indicators as the independent variable for regression analysis.

[0057] In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the method for constructing the user experience knowledge base of the present application will be described in detail below in conjunction with specific embodiments.

[0058] This embodiment relates to a specific method for constructing a user experience knowledge base, such as Figure 5 As shown, the following steps are included:

[0059] Based on the proposed multifaceted user experience model, a user experience knowledge base is constructed, which includes three important stages: user experience data discovery, user experience data integration, and user experience networking.

[0060] First, given a set of customer reviews as input, the user experience discovery process aims to discover terms or phrases that are highly relevant to the user experience through language unit recognition. Secondly, it collects business operation data from the background system of product services and extracts core data as key performance indicators to reflect the actual performance of products or services;

[0061] Using the extracted user experience data, the user experience integration process is activated to group the experience data based on different user experience aspects. In this step, several similarity functions are defined to measure the relationship between user experience data groupings under the same aspect, forming user experience data groups to represent different user experience aspects that customers care about.

[0062] User experience networking is based on user experience data groups and uses Bayesian networks to explore the relationships between various aspects of user experience. By calculating the probability of relationships between user experience aspects, the user experience network can help product service optimizers analyze user experience data and gain more insights into different user experiences. Finally, the identified user experience data, user experience aspect groups, and user experience networks together constitute the user experience knowledge base.

[0063] The embodiment of the present application also provides a device for constructing a user experience knowledge base. It should be noted that the device for constructing a user experience knowledge base in the embodiment of the present application can be used to execute the method for constructing a user experience knowledge base provided in the embodiment of the present application. The device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0064] The following is an introduction to the device for constructing a user experience knowledge base provided in an embodiment of the present application.

[0065] Figure 6 Schematic diagram of a device for constructing a user experience knowledge base according to an embodiment of the present application. Figure 6 As shown, the device comprises:

[0066] The acquisition unit 10 is used to acquire user experience data, wherein the user experience data is the reaction data generated during the interaction between the user and the product, and the user experience data includes user experience data, product performance data, product feature data and usage scenario data. The user experience data is used to characterize the subjective experience of the user in the process of using the product, the product performance data is used to characterize the performance index of the product in the process of the user using the product, the product feature data is used to characterize the functional features of the product in the process of the user using the product, and the usage scenario data is used to characterize the corresponding scenario in the process of the user using the product;

[0067] Specifically, user experience data can describe the corresponding psychological state of the product in a specific situation, such as positive, neutral and negative opinions. It can be expressed in emotional phrases, such as: easy to use, interesting. Product performance data can describe the key performance indicators of product services to reflect the actual operation of product services. These indicators indirectly affect customer satisfaction, such as: the proportion of automatic approval for credit card opening, the time from loan issuance to account receipt, the one-time pass rate of face recognition, etc. Product feature data refers to the product characteristics involved in user experience. It includes product entities, such as the credit card limit, rights and interests, and the appearance of the card. Usage scenario data can describe the scenario factors when using the product, such as location, time and activity.

[0068] A classification unit 20, configured to classify the plurality of user experience data according to the usage scenario data to obtain a plurality of data sets, wherein any one of the data sets includes a plurality of data groups, and any one of the data groups includes a plurality of user experience data;

[0069] Specifically, the user experience data integration process will be grouped according to different user experience aspects. In this solution, the focus is on grouping user experience data according to scenarios.

[0070] A construction unit 30 is used to construct a target user experience network based on the multiple data sets to obtain a user experience knowledge base, wherein the target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data;

[0071] Specifically, the target user experience network is expressed as: C is the product feature data, P is the product performance data, Q is the user experience data, and T is the usage scenario data.

[0072] The search unit 40 is used to obtain target search conditions, search in the user experience knowledge base according to the target search conditions, and obtain target search results, wherein the target search conditions are user experience data to be processed, and the target search results include the probability of any one of the product performance data, product feature data and usage scenario data affecting the user experience data to be processed.

[0073] Specifically, after the user experience knowledge base is built, it can be used to further optimize the customer experience. When a large number of user experience questionnaires are collected, it is impossible to analyze them one by one manually. The user experience knowledge base can be used to filter out questionnaires with negative emotions, and then experience optimization personnel can intervene to solve user problems. In addition, product business technicians will re-optimize products or business processes based on the user experience knowledge base. After the optimization, they will observe the user experience knowledge base again after a period of time to determine whether the user experience has improved.

[0074] Through this embodiment, a device for constructing a user experience knowledge base is provided, wherein an acquisition unit acquires user experience data, which is the reaction data generated during the interaction between a user and a product, and the user experience data includes user experience data, product performance data, product feature data, and usage scenario data; a classification unit classifies multiple user experience data according to the usage scenario data to obtain multiple data sets; a construction unit constructs a target user experience network according to the multiple data sets to obtain a user experience knowledge base, and the target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data; a search unit acquires a target search condition, searches in the user experience knowledge base according to the target search condition, and obtains a target search result. This solution reflects the key factors of user experience data by combining the four aspects of user experience data, product performance data, product feature data, and usage scenario data, so as to achieve the expansion of user experience data and reduce limitations; in addition, by classifying the user experience data and constructing the target user experience network, the construction of the user experience knowledge base is achieved, and then by searching the user experience knowledge base, the subjectivity of the causal dependency analysis between user experience data can be reduced, solving the problem that the user experience analysis in the prior art is relatively subjective.

[0075] In the specific implementation process, the acquisition unit includes an acquisition module, a first determination module, a first processing module, a second processing module, a third processing module, a second determination module, a third determination module, a fourth determination module, a fifth determination module, a sixth determination module and a seventh determination module, wherein the acquisition module is used to acquire user comment data and product usage data, wherein the user comment data includes the user's comment text on the product, and the product usage data includes the process data generated by the user in the process of using the product; the first determination module is used to determine the user experience data according to the user comment data; the first processing module is used to perform word segmentation on the comment text to obtain multiple initial words and initial phrases, wherein the comment text includes multiple initial sentences, and each of the initial sentences includes multiple initial words; the second processing module is used to perform part-of-speech tagging on the multiple initial words respectively to obtain the part-of-speech corresponding to each of the initial words, and perform part-of-speech tagging on the multiple initial phrases respectively to obtain the part-of-speech corresponding to each of the initial phrases, wherein the part-of-speech includes at least noun, verb, adjective and preposition; the second determination module is used to determine the part-of-speech corresponding to the verb and the part-of-speech corresponding to the adjective. The initial word corresponding to the adjective is the first preliminary opinion word, and the positive words whose number is in the top x among the multiple positive words, the synonyms of the positive words whose number is in the top x, the negative words whose number is in the top x among the multiple negative words, and the synonyms of the negative words whose number is in the top x are determined as the second preliminary opinion words, wherein x is a positive integer less than a predetermined value, and the predetermined value is the minimum value of the number of the positive words and the number of the negative words; the third determination module is used to determine that the same initial word in the first preliminary opinion word and the second preliminary opinion word is the target opinion word, and determine the target opinion word to be The above-mentioned initial sentence is the target opinion sentence; the fourth determination module is used to determine the above-mentioned initial word and the above-mentioned initial phrase corresponding to the above-mentioned noun as the first candidate word, and determine the above-mentioned preposition, the above-mentioned verb and the above-mentioned initial phrase corresponding to the above-mentioned noun as the second candidate word; the fifth determination module is used to determine the above-mentioned product feature data at least based on the above-mentioned target opinion sentence and the above-mentioned first candidate word; the sixth determination module is used to determine the above-mentioned usage scenario data at least based on the above-mentioned product feature data and the above-mentioned second candidate word; the seventh determination module is used to determine the above-mentioned product performance data based on the above-mentioned product usage data. The device can quickly obtain the above-mentioned user experience data, product performance data, product feature data and usage scenario data, and further improve the comprehensiveness of the above-mentioned user experience data.

[0076] Specifically, opinion words used to express subjective opinions in customer reviews are extracted. Adjectives are mostly used to express opinions, and verbs are usually used to convey customers' emotional orientation. Therefore, first, adjectives and verbs are extracted from the reviews as candidate opinion words using Part-Of-Speech tagging (POS). Then, based on the results of POS, the top n positive and negative words are manually selected from the adjectives and verbs in the reviews as seed word lists, respectively. Using WordNet, synonyms of seed words are added to expand, and finally a positive and negative word set is formed. The final opinion word is extracted by checking whether the candidate opinion word is included in the positive and negative word set, and the comment sentence with any opinion word is regarded as an opinion sentence. Then, given a set of user reviews as input, the user experience discovery process aims to discover terms or phrases that are highly related to the user experience through language unit recognition, namely user experience data, product feature data, and usage scenario data. Secondly, the business operation data is collected from the background system of the product service, and the core data is extracted as the key performance indicator, namely product performance data, to reflect the actual performance of the product or service.

[0077] In order to further quickly obtain the above-mentioned user perception data, the above-mentioned first determination module of the present application includes a first acquisition sub-module and an input sub-module, wherein the first acquisition sub-module is used to obtain historical user comment data and user perception data corresponding to the above-mentioned historical user comment data, wherein the above-mentioned historical user comment data is the above-mentioned user comment data obtained before the current moment, and the above-mentioned user perception data includes the emotional tendency of the above-mentioned user; the input sub-module is used to input multiple of the above-mentioned user comment data into the target neural network in sequence to obtain multiple user perception data corresponding to the above-mentioned user comment data, wherein the above-mentioned target neural network is trained by machine learning using multiple groups of historical data, and each group of data in the multiple groups of the above-mentioned historical data includes: the above-mentioned historical user comment data and user perception data corresponding to the above-mentioned historical user comment data.

[0078] Specifically, the process of obtaining the user experience data corresponding to the historical user comment data is as follows: first, use the Chinese word vector to convert the word text into a floating point vector, use jieba word segmentation to accurately segment the user comments, and normalize the length of each comment according to the normal distribution to cut off the long and make up the short. After the data preparation is completed, load the customer comment data, 90% of the samples are used for training, and 10% of the samples are used for testing. Build a neural network model for training. The first layer is the word embedding layer, which puts in the word vector matrix. The second layer uses a 64-unit bidirectional long short-term memory network, the third layer uses a 16-unit LSTM, and the last is the full connection layer, using the sigmoid activation function to output the result. Configure the training method, the loss function is the binary cross entropy loss function, the optimization function uses the Adam function, and the model evaluation function uses the accuracy function.

[0079] The fifth determination module includes a first calculation submodule, a second acquisition submodule, a first determination submodule, a third acquisition submodule, a second calculation submodule, a third calculation submodule and a sorting submodule, wherein the first calculation submodule is used to calculate the word frequency of the first candidate word in the user comment data to obtain the first word frequency; the second acquisition submodule is used to obtain the number of the target opinion sentences that do not include the first candidate word in the multiple target opinion sentences to obtain the first number; the first determination submodule is used to determine that the initial sentences that are not the target opinion sentences in the multiple initial sentences are non-target opinion sentences; the third acquisition submodule is used to obtain the number of non-target opinion sentences that include the first candidate word in the multiple non-target opinion sentences to obtain the second number; the second calculation submodule is used to calculate the distance between the initial sentence containing the first candidate word and the target opinion sentence to obtain the sentence distance; the third calculation submodule is used to calculate the distance between the initial sentence containing the first candidate word and the target opinion sentence according to the formula The weight w of the first candidate words is calculated, where t is the first word frequency, n1 is the first number, n2 is the second number, and d is the sentence distance; the sorting submodule is used to sort the weights of the first candidate words, and obtain the y first candidate words with the largest weights as the product feature data. The device can further quickly determine the product feature data.

[0080] Specifically, if users have more opinions on a particular product feature than others, this means that the feature has a greater impact on the user experience. Therefore, by defining the weight function for calculating the first candidate word as described above, the function assumes that if a noun or noun phrase is closer to any opinion word, that is, the smaller the function value, then it is more likely to become a product feature.

[0081] In some embodiments, the sixth determination module includes a second determination submodule, a fourth acquisition submodule, a fourth calculation submodule, and a third determination submodule, wherein the second determination submodule is used to determine that the initial sentence including the target opinion word and at least one of the first candidate words is a candidate sentence, and determine that the candidate sentence including the second candidate word is a target sentence; the fourth acquisition submodule is used to obtain any two of the second candidate words in the target sentence to obtain a candidate word group; the fourth calculation submodule is used to calculate the conditional symmetric probability of the two second candidate words in each candidate word group to obtain multiple conditional symmetric probabilities; the third determination submodule is used to determine that the second candidate word corresponding to the maximum value of the multiple conditional symmetric probabilities is the usage scenario data. The device can further quickly determine the usage scenario data.

[0082] Specifically, the selection function measures the phrase pair (v a , v b ) to calculate the score of fragment t. The higher the scp, the stronger the correlation. The calculation formula of conditional symmetric probability is as follows: Where n is v a and v b The length, w i Represents v a and v b The i-th word in the corpus, P(w1, w2) represents the probability of a given phrase appearing in the corpus. Then design some checking rules for screening, which use the language characteristics in or between segment groups to select segments. The above checking rules are as follows: check each phrase segment, if the number of words in the phrase segment is less than 2, delete it; group the phrase segments in the set according to whether they have the same phrase pair; in each group, select the phrases with prepositions or v a 、v b The phrase s with the smallest length is added to the set; if the phrase s j Contains phrase segments k , then remove s from the set j .

[0083] The seventh determination module includes a division submodule, a fifth acquisition submodule and a sixth acquisition submodule, wherein the division submodule is used to divide the process data to obtain multiple process nodes; the fifth acquisition submodule is used to obtain the contact points corresponding to each of the process nodes to obtain multiple contact points, wherein the contact points are used to characterize the interaction process between the product and the user; the sixth acquisition submodule is used to obtain the influencing factor data of each of the contact points to obtain the product performance data, wherein the influencing factor data is used to characterize the data that affects the user experience of the user. The device can further quickly obtain the product performance data.

[0084] Specifically, by collecting key performance indicators of products or services, the actual operation of the business can be reflected. Confirm the product service process, and split the entire service process into various links according to the service steps. In each link, analyze the contact points between customers and the company through face-to-face, online and other channels. Analyze the factors that may affect the customer experience at each contact point. In response to these influencing factors, collect data indicators that reflect the actual operation and performance of the business. Taking credit card services as an example, the key performance indicators that can be collected are as follows: the length of the credit card opening process, the time from the completion of loan approval to the actual receipt of the loan, the proportion of automatic approval, the one-time pass rate of facial recognition, the equipment failure rate, the customer service manual connection rate, the points redemption rate, etc. The optimization of these key performance indicators can improve customer experience satisfaction. The specific process of obtaining user experience data mentioned above is as follows: Figure 3 shown.

[0085] The classification unit includes a first calculation module and a second calculation module. The first calculation module is used to calculate the Calculate the local connection M l , where |s j ∩s k | is the jth usage scenario data s j and the kth usage scenario data s k The number of identical language fragments in |s j ∪s k | is the jth usage scenario data s j The number of language fragments and the k-th usage scenario data s k The second calculation module is used to calculate the sum of the number of language segments, the above-mentioned language segments include words, phrases and sentences; the second calculation module is used to calculate the sum of the number of language segments, the above-mentioned language segments include words, phrases and sentences according to the formula M g + = sim(r aj +r bk ), calculate the global connection M g , where r aj To include j The a-th user comment data, r bk To include k The b-th user comment data, sim(r aj +r bk ) is r aj and r bk similarity; Step S2023, according to the formula M = α × M l +β×M g , a connection matrix M is calculated, wherein α, β∈[0,1], α+β=1, and the connection matrix is ​​used to characterize the association relationship between the plurality of the data sets. The device can further realize the classification of user experience data quickly.

[0086] Specifically, the assumption of the above local connection is that if two context features s j and k Sharing more common language fragments indicates that they are more likely to be related to each other. For example, the scenario phrases "when selecting goods" and "when redeeming goods" are closely related to the context of consumers shopping in the points mall. The assumption of global connection is that if two scenario features are often mentioned in similar language environments, then they are likely to refer to similar aspects. For example, when talking about a specific product feature "credit card points", scenario phrases such as "in the mall" and "when redeeming" are usually accompanied by "points". The higher the frequency of occurrence, the higher the possibility that these scenario data are classified as user experience aspect groups related to "credit card points". In practical applications, text can be converted into word vectors, and then the cosine similarity between vectors can be calculated to represent the correlation between scenarios. The k-nearest neighbor algorithm can also be used to group the connection matrix to obtain the corresponding classification.

[0087] The construction unit includes a construction module, an eighth determination module, a third calculation module and a third processing module, wherein the construction module is used to construct an initial directed acyclic graph according to the data set, wherein the directed acyclic graph includes four data node groups, each of the data node groups includes a plurality of data nodes, and the data node groups correspond one to one with the user perception data, product performance data, product feature data and usage scenario data; the eighth determination module is used to determine whether any two of the user perception data, product feature data and usage scenario data are located in the same initial statement, and in the case where any two of the user perception data, product feature data and usage scenario data are located in the same initial statement, a connection is established between two corresponding data nodes, and the corresponding data nodes are any two corresponding data nodes of the user perception data, product feature data and usage scenario data located in the same initial statement; the third calculation module is used to calculate the corresponding data nodes according to the formula Calculate the value p of the above connection to update the above initial directed acyclic graph to obtain a preliminary user experience network, where z e is the eth corresponding data node, z f is the fth corresponding data node, |q(z e , z f )| is the number of initial statements in the above-mentioned initial statements that also include the user experience data corresponding to the above-mentioned corresponding data node, |q(z f )| for multiple of the above initial statements, including z fThe third processing module is used to process the product performance data and the user experience data using a multivariate linear regression analysis method to obtain the connection relationship between the data nodes corresponding to the product performance data and the data nodes corresponding to the user experience data, so as to update the preliminary user experience network and obtain the target user experience network. The device further quickly constructs the user experience knowledge base by constructing the target user experience network.

[0088] Specifically, Figure 4 As shown in the figure, there are four types of nodes in the user experience network, namely, scenario features, i.e., usage scenario data, product features, i.e., product feature data, user emotional state, i.e., user perception data, and key performance indicators, i.e., product performance data. An edge pointing from one node to another node represents the influence of one node on another node. In the three types of nodes, scenario features, product features, and user emotional state, the links between nodes are automatically formed based on text comments. If two nodes appear in the same sentence, an edge is assigned. Based on the influence on user experience, the directions of the edges include: from product features to scenario features; from product features to user emotional state; and from scenario features to user emotional state. In the two types of nodes, i.e., key performance indicators and user emotional state, the links between nodes are obtained through multivariate linear regression analysis, with user emotional state as the dependent variable and key performance indicators as the independent variable for regression analysis.

[0089] The device for constructing the user experience knowledge base includes a processor and a memory. The acquisition unit, classification unit, construction unit, and search unit are all stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions. The modules are all located in the same processor; or, the modules are located in different processors in any combination.

[0090] The processor includes a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and the user experience knowledge base can be built by adjusting kernel parameters.

[0091] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0092] An embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for constructing the user experience knowledge base.

[0093] An embodiment of the present invention provides a processor, and the processor is used to run a program, wherein the method for constructing the user experience knowledge base is executed when the program is running.

[0094] An embodiment of the present invention provides a device, the device includes a processor, a memory, and a program stored in the memory and executable on the processor, and when the processor executes the program, at least the steps of the method for constructing a user experience knowledge base are implemented. The device in this article can be a server, a PC, a PAD, a mobile phone, etc.

[0095] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes at least the steps of a method for constructing a user experience knowledge base.

[0096] Obviously, it should be understood by those skilled in the art that the modules or steps of the present invention described above can be implemented with a general-purpose computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented with executable program codes of computing devices, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the present invention is not limited to any specific combination of hardware and software. It should be understood by those skilled in the art that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0097] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0098] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory. Memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium. Computer-readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media, such as modulated data signals and carrier waves. It should also be noted that the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device that includes a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of further restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device that includes the elements.

[0099] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:

[0100] 1) The method for constructing the user experience knowledge base of the present application is as follows: first, user experience data is obtained. User experience data is the reaction data generated during the interaction between users and products. User experience data includes user experience data, product performance data, product feature data, and usage scenario data; then, multiple user experience data are classified according to the usage scenario data to obtain multiple data sets; then, a target user experience network is constructed according to the multiple data sets to obtain a user experience knowledge base. The target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data; finally, target search conditions are obtained, and the user experience knowledge base is searched according to the target search conditions to obtain target search results. This solution reflects the key factors of user experience data by combining four aspects of data, namely user experience data, product performance data, product feature data, and usage scenario data, so as to expand user experience data and reduce limitations; in addition, the construction of the user experience knowledge base is realized by classifying the user experience data and constructing the target user experience network, and then the subjectivity of the causal dependency analysis between user experience data can be reduced through the search of the user experience knowledge base, solving the problem that user experience analysis in the prior art is relatively subjective.

[0101] 2) The device for constructing the user experience knowledge base of the present application, the acquisition unit acquires user experience data, the user experience data is the reaction data generated during the interaction between the user and the product, and the user experience data includes user experience data, product performance data, product feature data and usage scenario data; the classification unit classifies multiple user experience data according to the usage scenario data to obtain multiple data sets; the construction unit constructs a target user experience network according to the multiple data sets to obtain a user experience knowledge base, and the target user experience network is used to characterize the probability of any one of the product performance data, product feature data and usage scenario data affecting the user experience data; the search unit acquires the target search condition, searches in the user experience knowledge base according to the target search condition, and obtains the target search result. This scheme reflects the key factors of user experience data by combining the four aspects of data, namely user experience data, product performance data, product feature data and usage scenario data, so as to achieve the expansion of user experience data and reduce limitations; in addition, the construction of the user experience knowledge base is achieved by classifying the user experience data and constructing the target user experience network, and then by searching the user experience knowledge base, the subjectivity of the causal dependency analysis between user experience data can be reduced, solving the problem that the user experience analysis in the prior art is relatively subjective.

[0102] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for constructing a user experience knowledge base, characterized in that: include: Obtaining user experience data, wherein the user experience data is reaction data generated during the interaction between the user and the product, and the user experience data includes user experience data, product performance data, product feature data, and usage scenario data. The user experience data is used to characterize the subjective experience of the user in the process of using the product, the product performance data is used to characterize the performance indicators of the product in the process of the user using the product, the product feature data is used to characterize the functional features of the product in the process of the user using the product, and the usage scenario data is used to characterize the corresponding scenario in the process of the user using the product; Classifying the plurality of user experience data according to the usage scenario data to obtain a plurality of data sets, wherein any one of the data sets includes a plurality of data groups, and any one of the data groups includes a plurality of the user experience data; According to the plurality of data sets, a target user experience network is constructed to obtain a user experience knowledge base, wherein the target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user perception data; Obtain target search conditions, search the user experience knowledge base according to the target search conditions, and obtain target search results, wherein the target search conditions are user experience data to be processed, and the target search results include the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data to be processed.

2. The method according to claim 1, characterized in that Obtain user experience data, including: Acquire user comment data and product usage data, wherein the user comment data includes the user's comment text on the product, and the product usage data includes process data generated by the user in the process of using the product; Determining the user experience data according to the user comment data; Performing word segmentation processing on the review text to obtain a plurality of initial words and initial phrases, wherein the review text includes a plurality of initial sentences, and each of the initial sentences includes a plurality of initial words; Performing part-of-speech tagging processing on the multiple initial words respectively to obtain the part-of-speech corresponding to each of the initial words, and performing part-of-speech tagging processing on the multiple initial phrases respectively to obtain the part-of-speech corresponding to each of the initial phrases, wherein the part-of-speech includes at least noun, verb, adjective and preposition; Determine the initial word corresponding to the verb as the part of speech and the initial word corresponding to the adjective as the first preliminary opinion word, and determine the positive words whose number is in the top x places among a plurality of positive words, the synonyms of the positive words whose number is in the top x places, the negative words whose number is in the top x places among a plurality of negative words, and the synonyms of the negative words whose number is in the top x places as the second preliminary opinion word, wherein x is a positive integer less than a predetermined value, and the predetermined value is the minimum value of the number of the positive words and the number of the negative words; determining that the same initial word in the first preliminary opinion word and the second preliminary opinion word is a target opinion word, and determining that the initial sentence corresponding to the target opinion word is a target opinion sentence; Determine the initial word corresponding to the noun and the initial phrase corresponding to the noun phrase as the first candidate word, and determine the initial phrase corresponding to the preposition, the verb and the noun as the second candidate word; determining the product feature data at least based on the target opinion sentence and the first candidate word; determining the usage scenario data at least based on the product feature data and the second candidate word; The product performance data is determined based on the product usage data.

3. The method according to claim 2, characterized in that Determining the user experience data according to the user comment data includes: Acquire historical user comment data and user feeling data corresponding to the historical user comment data, wherein the historical user comment data is the user comment data acquired before the current moment, and the user feeling data includes the emotional tendency of the user; Input the multiple user comment data into the target neural network in sequence to obtain the user perception data corresponding to the multiple user comment data, wherein the target neural network is trained by machine learning using multiple groups of historical data, and each group of the multiple groups of historical data includes: the historical user comment data and the user perception data corresponding to the historical user comment data.

4. The method according to claim 2, characterized in that: Determining the product feature data at least based on the target opinion statement and the first candidate word includes: Calculating the word frequency of the first candidate word in the user comment data to obtain a first word frequency; Obtaining the number of the target opinion sentences that do not include the first candidate word in the plurality of target opinion sentences to obtain a first number; Determining that the initial sentences that are not the target opinion sentences among the multiple initial sentences are non-target opinion sentences; Obtaining the number of non-target opinion sentences including the first candidate word from the plurality of non-target opinion sentences to obtain a second number; Calculating the distance between the initial sentence containing the first candidate word and the target opinion sentence to obtain a sentence distance; According to the formula Calculate the weight w of the first candidate word, where t is the first word frequency, n1 is the first number, n2 is the second number, and d is the sentence distance; The weights of the plurality of first candidate words are sorted, and y first candidate words with the largest weights are obtained as the product feature data.

5. The method according to claim 3, characterized in that: Determining the usage scenario data at least based on the product feature data and the second candidate word includes: determining the initial sentence including the target opinion word and at least one of the first candidate words as a candidate sentence, and determining the candidate sentence including the second candidate word as a target sentence; Obtain any two of the second candidate words from the plurality of the second candidate words in the target sentence to obtain a candidate word group; Calculating the conditional symmetric probabilities of two of the second candidate words in each of the candidate word groups to obtain a plurality of conditional symmetric probabilities; Determine that the second candidate word corresponding to the maximum value among the plurality of conditional symmetric probabilities is the usage scenario data.

6. The method according to claim 2, characterized in that Determining the product performance data according to the product usage data includes: Dividing the process data to obtain multiple process nodes; Acquire the contact points corresponding to each of the process nodes to obtain a plurality of contact points, wherein the contact points are used to characterize the interaction process between the product and the user; Obtain influencing factor data of each of the contact points to obtain the product performance data, wherein the influencing factor data is used to characterize data that affects the user experience of the user.

7. The method according to claim 1, characterized in that According to the usage scenario data, the plurality of user experience data are classified to obtain a plurality of data sets, including: According to the formula Calculate the local connection M l , where |s j ∩s k | is the jth usage scenario data s j and the kth usage scenario data s k The number of identical language fragments in |s j ∪s k | is the jth usage scenario data s j The number of language segments and the k-th usage scenario data s k The sum of the number of language segments, including words, phrases and sentences; According to the formula M g + = sim(r aj +r bk ), calculate the global connection M g , where r aj To include j The a-th user comment data, r bk To include k The b-th user comment data, sim(r aj +r bk ) is r aj and r bk similarity; According to the formula M = α × M l +β×M g , a connection matrix M is calculated, where α, β∈[0,1], α+β=1, and the connection matrix is ​​used to characterize the association relationship between the multiple data sets.

8. The method according to claim 2, characterized in that: According to the plurality of data sets, a target user experience network is constructed to obtain a user experience knowledge base, including: According to the data set, an initial directed acyclic graph is constructed, wherein the initial directed acyclic graph includes four data node groups, each of the data node groups includes a plurality of data nodes, and the data node groups correspond one to one with one of the user experience data, product performance data, product feature data, and usage scenario data; Determine whether any two of the user experience data, product feature data, and usage scenario data are located in the same initial statement. If any two of the user experience data, product feature data, and usage scenario data are located in the same initial statement, establish a connection between two corresponding data nodes, where the corresponding data nodes are any two corresponding data nodes of the user experience data, product feature data, and usage scenario data located in the same initial statement; According to the formula Calculate the value p of the connection to update the initial directed acyclic graph to obtain a preliminary user experience network, where z e is the corresponding data node of the eth node, z f is the corresponding data node of the fth node, |q(z e , z f )| is the number of initial statements in the plurality of initial statements that also include the user experience data corresponding to the corresponding data node, |q(z f )| for multiple initial statements, including z f The number of initial sentences of the corresponding user experience data; The product performance data and the user experience data are processed by a multivariate linear regression analysis method to obtain the connection relationship between the data nodes corresponding to the product performance data and the data nodes corresponding to the user experience data, so as to update the preliminary user experience network and obtain the target user experience network.

9. A device for constructing a user experience knowledge base, characterized in that: include: an acquisition unit, configured to acquire user experience data, wherein the user experience data is response data generated during the interaction between a user and a product, and includes user experience data, product performance data, product feature data, and usage scenario data, wherein the user experience data is used to characterize the subjective feelings of the user during the process of using the product, the product performance data is used to characterize the performance indicators of the product during the process of the user using the product, the product feature data is used to characterize the functional features of the product during the process of the user using the product, and the usage scenario data is used to characterize the corresponding scenario during the process of the user using the product; A classification unit, configured to classify the plurality of user experience data according to the usage scenario data to obtain a plurality of data sets, wherein any one of the data sets includes a plurality of data groups, and any one of the data groups includes a plurality of the user experience data; A construction unit, configured to construct a target user experience network based on the plurality of data sets to obtain a user experience knowledge base, wherein the target user experience network is used to characterize the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user perception data; A search unit is used to obtain a target search condition, search the user experience knowledge base according to the target search condition, and obtain a target search result, wherein the target search condition is the user experience data to be processed, and the target search result includes the probability of any one of the product performance data, product feature data, and usage scenario data affecting the user experience data to be processed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 8.

Images