Data generation method, device, medium and electronic device

Abstract

The present invention provides a data generation method, device, medium and electronic device. The data generation method provided by the embodiment of the present invention includes: obtaining historical evaluation data of an object to be evaluated, and extracting evaluation keywords that can be used to characterize the physical characteristics of the object to be evaluated from the historical evaluation data, and then obtaining evaluation data by inputting the evaluation keywords into a preset neural network model, wherein the evaluation data can be used as the core evaluation data of the object to be evaluated and displayed in the first preset position in the evaluation area of ​​the object to be evaluated. The data generation method provided by the embodiment of the present invention can output high-value evaluation data for the user to characterize the relevant characteristics of the object to be evaluated.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a data generation method, device, medium and electronic equipment. Background Art

[0002] With the rise of e-commerce, online shopping has become a popular way of shopping with two major advantages: convenience and product diversity.

[0003] Among them, in the process of online shopping, there is a crisis of trust in the products, that is, the difference between the limited product knowledge obtained by users through browsing the product pictures or text descriptions on the application (APP) or website and the actual products received. Therefore, when users are satisfied with the characteristics of the product (appearance, material, price), they will browse the comments of the purchased users on the product to make further decisions.

[0004] However, since product review information is usually large in quantity and the content is confusing, users who view product reviews cannot efficiently obtain the core evaluation data of the product. Summary of the invention

[0005] The embodiments of the present invention provide a data generation method, device, medium and electronic device to provide users with high-value commodity data and provide a reliable reference for users' shopping decisions.

[0006] In a first aspect, an embodiment of the present invention provides a data generation method, including:

[0007] Acquire historical evaluation data of the object to be evaluated, and extract evaluation keywords of the historical evaluation data, wherein the evaluation keywords are used to characterize the physical properties of the object to be evaluated;

[0008] Generate evaluation data corresponding to the evaluation keyword according to the evaluation keyword and a preset neural network model;

[0009] The evaluation data is displayed at a first preset position in the evaluation area of ​​the object to be evaluated.

[0010] In a possible design, after displaying the evaluation data at a preset position in the evaluation area of ​​the object to be evaluated, the method further includes:

[0011] Acquire a first evaluation instruction input for first evaluation data, wherein the evaluation data includes the first evaluation data;

[0012] The first sub-evaluation corresponding to the first evaluation instruction is displayed at a second preset position in the evaluation area of ​​the object to be evaluated, the first sub-evaluation is used to evaluate the first evaluation data, and the second preset position is adjacent to the first preset position in the evaluation area.

[0013] In a possible design, extracting evaluation keywords of the historical evaluation data includes:

[0014] Performing word segmentation processing on the historical evaluation data according to a preset word library to obtain a word segmentation set;

[0015] Calculate the weight value of each word in the word set according to a preset keyword weight algorithm;

[0016] The word segment whose weight value is greater than the preset weight value is selected as the first evaluation keyword of the historical evaluation data, and the evaluation keywords include the first evaluation keyword.

[0017] In a possible design, the step of obtaining historical evaluation data of the object to be evaluated includes:

[0018] When the number of evaluation items corresponding to the object to be evaluated is greater than a preset entry number threshold, the current evaluation is obtained as the historical evaluation data corresponding to the object to be evaluated.

[0019] In a possible design, before performing word segmentation processing on the historical evaluation data according to the preset word library, the method further includes:

[0020] The historical evaluation data is preprocessed, and the preprocessing includes: removing stop words and punctuation marks in the historical evaluation data.

[0021] In a possible design, the obtaining of historical evaluation data of the object to be evaluated and extracting evaluation keywords of the historical evaluation data includes:

[0022] When the evaluation item corresponding to the object to be evaluated is less than or equal to the preset item threshold, the evaluation keyword corresponding to the third-level category corresponding to the object to be evaluated is obtained as the second evaluation keyword, and the evaluation keyword includes the second evaluation keyword.

[0023] In a possible design, generating evaluation data corresponding to the evaluation keyword according to the evaluation keyword and a preset neural network model includes:

[0024] Generate first evaluation data corresponding to the object to be evaluated according to the first evaluation keyword and a preset neural network model;

[0025] Generate second evaluation data corresponding to the object to be evaluated according to the first evaluation keyword and a preset neural network model;

[0026] The evaluation data includes the first evaluation data and the second evaluation data.

[0027] In a possible design, before generating the evaluation data corresponding to the evaluation keyword according to the evaluation keyword and the preset neural network model, the method further includes:

[0028] The preset neural network model is trained according to the evaluation keywords and the historical evaluation data, and the preset neural network model is a long short-term memory network model.

[0029] In a possible design, the training of the preset neural network model according to the evaluation keywords includes:

[0030] The evaluation keywords and each word in the word set are vectorized, and the vectorized data matrix is ​​used as training data input to train the long short-term memory network model.

[0031] In a second aspect, an embodiment of the present invention further provides a data generating device, including:

[0032] An acquisition module is used to obtain historical evaluation data of the object to be evaluated;

[0033] An extraction module, used for extracting evaluation keywords of the historical evaluation data, wherein the evaluation keywords are used for characterizing the physical properties of the object to be evaluated;

[0034] A processing module, used to generate evaluation data corresponding to the evaluation keywords according to the evaluation keywords and a preset neural network model;

[0035] A display module is used to display the evaluation data at a first preset position in the evaluation area of ​​the object to be evaluated.

[0036] In a possible design, the acquisition module is further used to acquire a first evaluation instruction input for first evaluation data, and the evaluation data includes the first evaluation data;

[0037] The display module is further used to display a first sub-evaluation corresponding to the first evaluation instruction at a second preset position in the evaluation area of ​​the object to be evaluated, the first sub-evaluation being used to evaluate the first evaluation data, and the second preset position being adjacent to the first preset position in the evaluation area.

[0038] In a possible design, the extraction module is specifically used to:

[0039] Performing word segmentation processing on the historical evaluation data according to a preset word library to obtain a word segmentation set;

[0040] Calculate the weight value of each word in the word set according to a preset keyword weight algorithm;

[0041] The word segment whose weight value is greater than the preset weight value is selected as the first evaluation keyword of the historical evaluation data, and the evaluation keywords include the first evaluation keyword.

[0042] In a possible design, the acquisition module is specifically used to:

[0043] When the number of evaluation items corresponding to the object to be evaluated is greater than a preset entry number threshold, the current evaluation is obtained as the historical evaluation data corresponding to the object to be evaluated.

[0044] In a possible design, the processing module is further used to preprocess the historical evaluation data, and the preprocessing includes: removing stop words and punctuation marks in the historical evaluation data.

[0045] In a possible design, the extraction module is specifically used to:

[0046] When the evaluation item corresponding to the object to be evaluated is less than or equal to the preset item threshold, the evaluation keyword corresponding to the third-level category corresponding to the object to be evaluated is obtained as the second evaluation keyword, and the evaluation keyword includes the second evaluation keyword.

[0047] In a possible design, the processing module is specifically used to:

[0048] Generate first evaluation data corresponding to the object to be evaluated according to the first evaluation keyword and a preset neural network model;

[0049] Generate second evaluation data corresponding to the object to be evaluated according to the first evaluation keyword and a preset neural network model;

[0050] The evaluation data includes the first evaluation data and the second evaluation data.

[0051] In one possible design, the data generating device further includes:

[0052] A training module is used to train the preset neural network model according to the evaluation keywords and the historical evaluation data, and the preset neural network model is a long short-term memory network model.

[0053] In a possible design, the training module is specifically used to:

[0054] The evaluation keywords and each word in the word set are vectorized, and the vectorized data matrix is ​​used as training data input to train the long short-term memory network model.

[0055] In a third aspect, an embodiment of the present invention further provides an electronic device, including:

[0056] processor; and,

[0057] A memory, configured to store executable instructions of the processor;

[0058] The processor is configured to execute any possible data generation method in the first aspect by executing the executable instructions.

[0059] In a fourth aspect, an embodiment of the present invention further provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements any possible data generation method in the first aspect.

[0060] A data generation method, device, medium and electronic device provided by an embodiment of the present invention obtains historical evaluation data of an object to be evaluated, and extracts evaluation keywords that can be used to characterize the physical characteristics of the object to be evaluated from the historical evaluation data, and then obtains evaluation data by inputting the evaluation keywords into a preset neural network model, wherein the evaluation data can be used as core evaluation data of the object to be evaluated and displayed in a first preset position in an evaluation area of ​​the object to be evaluated, thereby outputting high-value evaluation data for the user to characterize relevant characteristics of the object to be evaluated. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0062] Figure 1 is a schematic diagram of an application scenario architecture of a data generation method according to an exemplary embodiment of the present invention;

[0063] Figure 2 is a flow chart of a data generation method according to an exemplary embodiment of the present invention;

[0064] Figure 3 yes Figure 2 A schematic diagram of a possible product information browsing interface in the illustrated embodiment;

[0065] Figure 4is a flow chart of a data generation method according to another exemplary embodiment of the present invention;

[0066] Figure 5 yes Figure 4 A schematic diagram of a possible product evaluation interface in the illustrated embodiment;

[0067] Figure 6 is a schematic diagram of a preset neural network model training process according to an exemplary embodiment of the present invention;

[0068] Figure 7 is a structural schematic diagram of a data generating device according to an exemplary embodiment of the present invention;

[0069] Figure 8 is a structural schematic diagram of a data generating device according to another exemplary embodiment of the present invention;

[0070] Fig. 9 It is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0073] Generally speaking, the process of a user shopping online can be roughly simplified as follows: (1) unconsciously browsing products or searching for a product purposefully; (2) viewing product details; (3) viewing product reviews; (4) placing an order and paying for the product. In the above step (2), it is assumed that the user is relatively satisfied with the characteristics of the product (appearance, material texture, price). At this time, the user will not directly enter the purchase stage as when shopping in a physical store, because the user understands that the item seen at this time may be different from the real item. Therefore, the user will generally carefully browse the comments of users who have purchased the product and focus on some characteristics that they care about, such as color, style, etc. In this process, product review information largely determines the user's order purchase behavior.

[0074] At present, the product review information of online shopping products is voluntarily filled in by users after completing the transaction. Through a simple analysis of users' review behavior, we can use the product reviews on the JD Mall APP as a reference. In order to encourage users to provide review information, JD Mall uses the incentive strategy of giving JD beans for reviewing orders.

[0075] Specifically, in this scenario, users can be roughly divided into three categories: the first category is those who do not participate in commenting, the second category is real commentators, and the third category is those who comment for the Jingdou reward.

[0076] It is understandable that among the first category of people, some have no awareness of evaluation, and some have awareness of evaluation but think that writing reviews is a time-consuming and laborious task; the second category of people mainly comes from people who will comment even without any reward. These users have a strong awareness of independent evaluation and generally have clear likes and dislikes for the purchased products and want to share them with subsequent buyers. Therefore, the review information submitted by these people is more valuable. Some of them are people who evaluate seriously due to the stimulation of reward strategies; the third category of people are basically those who comment casually for the sake of rewards, and their comments are generally meaningless.

[0077] Moreover, the content of the product reviews filled in by users varies from person to person. Some reviews are concise and highlight the key points, while others are irrelevant. Therefore, too much product review information will cause trouble for users when making purchase decisions, and they will not be able to quickly obtain effective evaluation data.

[0078] In response to the above-mentioned problems, an embodiment of the present invention provides a data generation method, which obtains historical evaluation data of an object to be evaluated, extracts evaluation keywords that can be used to characterize the physical characteristics of the object to be evaluated from the historical evaluation data, and then obtains product evaluation data by inputting the evaluation keywords into a preset neural network model, wherein the product evaluation data can be used as the core evaluation data of the object to be evaluated, and is displayed in the first preset position in the evaluation area of ​​the object to be evaluated, thereby providing users with high-value product evaluation data to provide a reliable reference for the user's shopping decision. Optionally, the above-mentioned object to be evaluated can be understood as the product to be evaluated, and the above-mentioned historical evaluation data can be understood as the historical evaluation information of the product.

[0079] The device implementing the data generation method in this solution can be any type of electronic device that can play games, including tablet computers, smart phones, personal computers, etc. The data generation method is described in detail below through several specific implementations.

[0080] Figure 1 FIG. 1 is a schematic diagram of an application scenario architecture of a data generation method according to an exemplary embodiment of the present invention. Figure 1 As shown, when it is necessary to provide evaluation data for the object to be evaluated, for example, when it is necessary to provide product evaluation data for the product to be evaluated, a request can be sent through the gateway, and then the corresponding core evaluations can be obtained from the core evaluation database through the product evaluation acquisition service, and these core evaluations can be used as the product evaluation data of the product to be evaluated, and the response output can be displayed in the evaluation area of ​​the product to be evaluated.

[0081] The product review acquisition service may be obtained from a core review database through a product review acquisition interface, or may be obtained from a historical review acquisition interface based on historical review information already available for the current product to be evaluated.

[0082] It is worth understanding that, for the historical review information obtained from the historical review acquisition interface based on the current historical review information of the product to be evaluated, it can be when the product to be evaluated already has a certain amount of historical evaluation data. For example, a threshold value for the number of evaluations can be set. When the number of evaluations of the product exceeds the threshold, the evaluation keywords of the product to be evaluated can be extracted, that is, all the evaluation data can be aggregated into multiple evaluation sets, for example, each evaluation set has 100 evaluations, and then input into the trained long short-term memory network model (Long Short-Term Memory, referred to as LSTM) to generate evaluation data, and a mapping relationship can also be established between the evaluation data and the encoding of the product to be evaluated for subsequent query.

[0083] It is also worth understanding that some products have just been put on the shelves and have no historical review information or too little review information. At this time, the review keywords of the third-level category are used as the review keywords of the product to be evaluated. In this case, the core review database can be generated by the product evaluation generation service, and the product evaluation generation service can be online or offline. Specifically, it is to first obtain product information and historical evaluation data of the third-level category of the corresponding product to be evaluated, and then obtain the evaluation keywords in the historical evaluation data through a keyword weight algorithm, such as the term frequency-inverse document frequency index algorithm (Term Frequency-Inverse Document Frequency, referred to as TF-IDF) to generate an evaluation keyword database, and then input the evaluation keywords into the LSTM to generate evaluation data, and can also establish a mapping relationship between the product evaluation data and the third-level category for subsequent queries.

[0084] In addition, the evaluation keywords obtained by the above two methods can be input into LSTM to generate product evaluation data corresponding to the evaluation keywords. In addition, the product evaluation data is saved in the database with the mapping relationship of (code of the object to be evaluated: core evaluation 1, core evaluation 2...) and (third-level category: core evaluation 1, core evaluation 2...) for online query.

[0085] Figure 2 FIG. 1 is a flow chart of a data generation method according to an exemplary embodiment of the present invention. Figure 2 As shown, the data generation method provided in this embodiment includes:

[0086] Step 101: Obtain historical evaluation data of the object to be evaluated.

[0087] Specifically, when the number of evaluation items corresponding to the object to be evaluated is greater than a preset item number threshold, the current product evaluation is obtained as the historical evaluation data corresponding to the object to be evaluated. For example, when the number of evaluation data items corresponding to the object to be evaluated is greater than 100, these product evaluation data can be used as historical evaluation data to extract evaluation keywords. Among them, all historical evaluation data can also be extracted by aggregating them into multiple evaluation data sets, for example, each evaluation data set can include 100 comments.

[0088] Step 102: Extract evaluation keywords from historical evaluation data.

[0089] When the number of evaluation items corresponding to the product to be evaluated is greater than the preset entry number threshold, after obtaining the historical evaluation data of the object to be evaluated, the evaluation keywords of the historical evaluation data can be further extracted, wherein the evaluation keywords are used to characterize the physical properties of the product to be evaluated.

[0090] Specifically, the historical evaluation data may be segmented according to a preset word library to obtain a segmented word set, and then the weight value of each segmented word in the segmented word set is calculated according to a preset keyword weight algorithm, and the segmented word with a weight value greater than the preset weight value is selected as the first evaluation keyword of the historical evaluation data, and the evaluation keyword includes the first evaluation keyword. Optionally, in order to make the segmentation more accurate, the historical evaluation data may be preprocessed before the segmentation is performed on the historical evaluation data according to the preset word library, wherein the preprocessing may include: removing stop words and punctuation marks in the historical evaluation data.

[0091] It is worth noting that, in the present embodiment, the preset word library can be the Jieba word library or other word library, or it can be an extended word library with a custom dictionary added to the existing word library. And, the keyword weight algorithm can be the TF-IDF algorithm. Among them, the principle of the TF-IDF algorithm is: it actually calculates the product of TF and IDF first, and then uses the result of the product to measure the importance of the words in a word library to each document, so as to evaluate the importance of keywords to a file set or one of the files in a corpus. Among them, the specific implementation principle of the TF-IDF algorithm is an existing algorithm, which is not specifically limited in the present embodiment.

[0092] In addition, in another possible situation, when some products have just been put on the shelves and have no historical evaluation data or too little evaluation data, that is, when the evaluation item corresponding to the product to be evaluated is less than or equal to the preset item threshold, the evaluation keyword corresponding to the third-level category corresponding to the product to be evaluated can be obtained as the second evaluation keyword, and the evaluation keywords include the second evaluation keywords.

[0093] Step 103: Generate evaluation data corresponding to the evaluation keywords based on the evaluation keywords and the preset neural network model.

[0094] After obtaining the evaluation keywords, the evaluation keywords can be input into the trained LSTM model to generate evaluation data corresponding to the evaluation keywords. Among them, the training of the preset neural network model, such as the LSTM model, is described in detail in the subsequent embodiments. Among them, the model can be trained by deep learning. Deep learning is a type of machine learning algorithm, and deep learning can be understood as a data feature learning method. It uses a cascade of multiple layers of nonlinear processing units for feature extraction and conversion. Each continuous layer uses the output of the previous layer as input. By abstracting the features layer by layer, more advanced data features can be learned. LSTM is a special recurrent neural network. LSTM solves the problem that traditional recurrent neural networks are difficult to handle long-distance dependencies by adding input thresholds, forgetting thresholds, and output thresholds.

[0095] Step 104: Display the product evaluation data at a first preset position in the evaluation area of ​​the object to be evaluated.

[0096] After generating evaluation data corresponding to the evaluation keywords according to the evaluation keywords and the preset neural network model, the product evaluation data can also be displayed in the first preset position in the evaluation area of ​​the product to be evaluated to evaluate the product to be evaluated.

[0097] In one possible scenario, Figure 3 yes Figure 2 A schematic diagram of a possible product information browsing interface in the embodiment shown is shown. A mobile phone can be selected as an example of a product to be evaluated. Figure 3 As shown, the first preset position in the evaluation area can be Figure 3 The core comment area shown in the figure, and the product evaluation data generated by the method of this embodiment can be, for example: fully functional, comfortable to use, and perfect operation, etc. The product evaluation data in the core comment area can be the core evaluation data determined based on the historical evaluation data in all comment areas, or can be the core evaluation data corresponding to the third-level category of the product (mobile phone).

[0098] In this embodiment, historical evaluation data of the product to be evaluated is obtained, and physical property evaluation keywords that can be used to characterize the product to be evaluated are extracted from the historical evaluation data. Then, the product evaluation data is obtained by inputting the evaluation keywords into a preset neural network model. The product evaluation data can be used as the core evaluation data of the product to be evaluated and displayed in the first preset position in the evaluation area of ​​the product to be evaluated, thereby outputting high-value evaluation data for the user to characterize the relevant characteristics of the object to be evaluated.

[0099] Figure 4 FIG. 1 is a flow chart of a data generation method according to another exemplary embodiment of the present invention. Figure 4 As shown, the data generation method provided in this embodiment includes:

[0100] Step 201: Obtain historical evaluation data of the object to be evaluated.

[0101] Step 202: extract evaluation keywords from historical evaluation data.

[0102] Step 203: Generate evaluation data corresponding to the evaluation keywords according to the evaluation keywords and the preset neural network model.

[0103] Step 204: Display the product evaluation data at a first preset position in the evaluation page of the object to be evaluated.

[0104] It is worth noting that the specific implementation of steps 201-204 in this embodiment refers to Figure 2The description of steps 101-104 in the illustrated embodiment will not be repeated here.

[0105] Step 205: Obtain a first evaluation instruction for the first evaluation data input.

[0106] After the product evaluation data is displayed at the first preset position in the evaluation page of the product to be evaluated, in order to make it more convenient for users to evaluate the product, especially to increase the frequency of evaluation by the first type of users and to improve the credibility of the evaluation data of the third type of users. In addition to the traditional text, picture or video evaluation, the first evaluation instruction input for the first product evaluation data can also be obtained. Figure 3 , a simple and quick evaluation of the product to be evaluated can be made by "thumbs up" or "thumbs down", which not only simplifies the evaluation process of the user, but also effectively evaluates the overall quality of the product. It is worth understanding that in this embodiment, the first evaluation instruction can be the user's "thumbs up" or "thumbs down" evaluation input.

[0107] Step 206: Display the first sub-evaluation corresponding to the first evaluation instruction at a second preset position in the evaluation area of ​​the object to be evaluated.

[0108] Figure 5 yes Figure 4 A schematic diagram of a possible product evaluation interface in the embodiment shown. Figure 5 As shown, when evaluating the product to be evaluated, the data generation method provided by this embodiment can perform simple "like" or "dislike" operations on the core evaluation, and can also enter specific evaluations to conduct more dimensional and detailed evaluations on the evaluation object, which is conducive to establishing the credibility and value of the product evaluation.

[0109] In addition, the first sub-evaluation corresponding to the first evaluation instruction can also be displayed at a second preset position in the evaluation area of ​​the product to be evaluated, the first sub-evaluation is used to evaluate the first product evaluation data, and the second preset position is set adjacent to the first preset position in the evaluation area.

[0110] Figure 6 FIG. 1 is a schematic diagram of a preset neural network model training process according to an exemplary embodiment of the present invention. Figure 6 As shown, in any of the above embodiments, the process of training the preset neural network model includes:

[0111] Step 301: Obtain historical evaluation data of the object to be evaluated.

[0112] The historical evaluation data of the product to be evaluated may be obtained through the information acquisition interface, wherein the historical evaluation data may also be aggregated into multiple evaluation data sets.

[0113] Step 302: pre-process the historical evaluation data.

[0114] After obtaining the historical evaluation data of the product to be evaluated, the historical evaluation data may also be preprocessed, for example, by removing stop words and punctuation marks in the historical evaluation data.

[0115] Step 303: perform word segmentation processing on the historical evaluation data according to a preset word library.

[0116] Then, the historical evaluation data may be segmented according to a preset word library, wherein the preset word library may be the Jieba word library or other word library, or an extended word library with a custom dictionary added to an existing word library.

[0117] Step 304: extract evaluation keywords from historical evaluation data.

[0118] Step 305: vectorize the evaluation keywords.

[0119] The weight value of each word in the word set can be calculated according to a preset keyword weight algorithm, such as the TF-IDF algorithm, and the word with a weight value greater than the preset weight value is selected as the evaluation keyword of the historical evaluation data. After determining the evaluation keyword, it is vectorized. Among them, the vectorization method can be with the help of Word2vec. Specifically, Word2vec is a group of related models used to generate word vectors. These models are shallow and two-layer neural networks that are used for training to reconstruct linguistic word texts. The network is represented by words, and the input words in adjacent positions need to be guessed. Under the assumption of the bag of words model in word2vec, the order of words is not important. After training, the word2vec model can be used to map each word to a vector, which can be used to represent the relationship between words. The vector is the hidden layer of the neural network.

[0120] Step 306: vectorize each word in the word set.

[0121] In addition, each word in the word set can be vectorized, and the specific vectorization method can also be through Word2vec.

[0122] Step 307: Obtain a vectorized data matrix.

[0123] After the evaluation keywords are vectorized and each word in the word set is vectorized, the vectorized data is constructed into a data matrix.

[0124] Step 308: input to the long short-term memory network model for training.

[0125] The constructed data matrix is ​​input into the long short-term memory network model for training.

[0126] Figure 7 FIG. 1 is a schematic diagram of a data generating device according to an exemplary embodiment of the present invention. Figure 7 As shown, the data generating device provided in this embodiment includes:

[0127] The acquisition module 401 is used to acquire the historical evaluation data of the object to be evaluated;

[0128] An extraction module 402 is used to extract evaluation keywords of the historical evaluation data, where the evaluation keywords are used to characterize the physical properties of the object to be evaluated;

[0129] A processing module 403 is used to generate evaluation data corresponding to the evaluation keyword according to the evaluation keyword and a preset neural network model;

[0130] The display module 404 is used to display the evaluation data at a first preset position in the evaluation area of ​​the object to be evaluated, so as to evaluate the object to be evaluated.

[0131] In a possible design, the acquisition module 401 is further used to acquire a first evaluation instruction input for first evaluation data, and the evaluation data includes the first evaluation data;

[0132] The display module 404 is further used to display the first sub-evaluation corresponding to the first evaluation instruction at a second preset position in the evaluation area of ​​the object to be evaluated, the first sub-evaluation being used to evaluate the first evaluation data, and the second preset position is adjacent to the first preset position in the evaluation area.

[0133] In a possible design, the extraction module 402 is specifically used to:

[0134] Performing word segmentation processing on the historical evaluation data according to a preset word library to obtain a word segmentation set;

[0135] Calculate the weight value of each word in the word set according to a preset keyword weight algorithm;

[0136] The word segment whose weight value is greater than the preset weight value is selected as the first evaluation keyword of the historical evaluation data, and the evaluation keywords include the first evaluation keyword.

[0137] In a possible design, the acquisition module 401 is specifically used to:

[0138] When the number of evaluation items corresponding to the object to be evaluated is greater than a preset item number threshold, the current product evaluation is obtained as the historical evaluation data corresponding to the object to be evaluated.

[0139] In a possible design, the processing module 403 is further used to preprocess the historical evaluation data, and the preprocessing includes: removing stop words and punctuation marks in the historical evaluation data.

[0140] In a possible design, the extraction module 402 is specifically used to:

[0141] When the evaluation item corresponding to the object to be evaluated is less than or equal to the preset item threshold, the evaluation keyword corresponding to the third-level category corresponding to the object to be evaluated is obtained as the second evaluation keyword, and the evaluation keyword includes the second evaluation keyword.

[0142] In a possible design, the processing module 403 is specifically configured to:

[0143] Generate first evaluation data corresponding to the object to be evaluated according to the first evaluation keyword and a preset neural network model;

[0144] Generate second evaluation data corresponding to the object to be evaluated according to the first evaluation keyword and a preset neural network model;

[0145] The evaluation data includes the first evaluation data and the second evaluation data.

[0146] exist Figure 7 Based on the embodiment shown, Figure 8 FIG. 1 is a schematic diagram of a data generating device according to another exemplary embodiment of the present invention. Figure 8 As shown, the data generating device provided in this embodiment further includes:

[0147] The training module 405 is used to train the preset neural network model according to the evaluation keywords and the historical evaluation data, and the preset neural network model is a long short-term memory network model.

[0148] In a possible design, the training module 405 is specifically used to:

[0149] The evaluation keywords and each word in the word set are vectorized, and the vectorized data matrix is ​​used as training data input to train the long short-term memory network model.

[0150] The above processing module 403 can be configured as one or more integrated circuits for implementing the above method, such as: one or more application specific integrated circuits (ASIC), or one or more digital singnal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a certain module above is implemented in the form of a processing element scheduling program code, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call program codes. For another example, these modules can be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0151] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional units. It is worth noting that Figure 7-Figure 8 The data generation device provided in the illustrated embodiment can be used to execute the data generation method provided in any of the above method embodiments. The specific implementation method and technical effects are similar and will not be repeated here.

[0152] Fig. 9 FIG. 1 is a schematic diagram of the structure of an electronic device according to an exemplary embodiment of the present invention. Fig. 9 As shown, this embodiment provides an electronic device 500, including:

[0153] processor 501; and,

[0154] A memory 502, used to store executable instructions of the processor, and the memory may also be a flash memory;

[0155] The processor 501 is configured to execute the various steps in the above method by executing the executable instructions. For details, please refer to the relevant description in the above method embodiment.

[0156] Optionally, the memory 502 may be independent or integrated with the processor 501 .

[0157] When the memory 502 is a device independent of the processor 501, the electronic device may further include:

[0158] The bus 503 is used to connect the processor 501 and the memory 502 .

[0159] This embodiment further provides a readable storage medium, in which a computer program is stored. When at least one processor of an electronic device executes the computer program, the electronic device executes the methods provided in the above-mentioned various implementation modes.

[0160] This embodiment also provides a program product, which includes a computer program stored in a readable storage medium. At least one processor of the electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device implements the methods provided in the above various embodiments.

[0161] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A data generation method, It is characterized in that include: Acquire historical evaluation data of the object to be evaluated, and extract evaluation keywords of the historical evaluation data, wherein the evaluation keywords are used to characterize the physical properties of the object to be evaluated; Generate evaluation data corresponding to the evaluation keyword according to the evaluation keyword and a preset neural network model; The evaluation data is the core evaluation data of the object to be evaluated corresponding to each of the evaluation keywords; Displaying the evaluation data at a first preset position in the evaluation area of ​​the object to be evaluated; Acquire a first evaluation instruction input for first evaluation data, wherein the evaluation data includes the first evaluation data; The first sub-evaluation corresponding to the first evaluation instruction is displayed at a second preset position in the evaluation area of ​​the object to be evaluated, the first sub-evaluation is used to evaluate the first evaluation data, and the second preset position is adjacent to the first preset position in the evaluation area.

2. The data generation method according to claim 1, It is characterized in that The step of extracting evaluation keywords from the historical evaluation data includes: Performing word segmentation processing on the historical evaluation data according to a preset word library to obtain a word segmentation set; Calculate the weight value of each word in the word set according to a preset keyword weight algorithm; The word segment whose weight value is greater than the preset weight value is selected as the first evaluation keyword of the historical evaluation data, and the evaluation keywords include the first evaluation keyword.

3. The data generation method according to claim 2, It is characterized in that The obtaining of historical evaluation data of the object to be evaluated includes: When the number of evaluation items corresponding to the object to be evaluated is greater than a preset entry number threshold, the current evaluation is obtained as the historical evaluation data corresponding to the object to be evaluated.

4. The data generation method according to claim 3, It is characterized in that Before the word segmentation processing is performed on the historical evaluation data according to the preset word library, the method further includes: The historical evaluation data is preprocessed, and the preprocessing includes: removing stop words and punctuation marks in the historical evaluation data.

5. The data generation method according to claim 2, It is characterized in that The step of obtaining historical evaluation data of the object to be evaluated and extracting evaluation keywords of the historical evaluation data includes: When the number of evaluation items corresponding to the object to be evaluated is less than or equal to a preset item number threshold, the evaluation keyword corresponding to the third-level category corresponding to the object to be evaluated is obtained as the second evaluation keyword, and the evaluation keyword includes the second evaluation keyword.

6. The data generation method according to claim 5, It is characterized in that The step of generating evaluation data corresponding to the evaluation keyword according to the evaluation keyword and a preset neural network model includes: Generate first evaluation data corresponding to the object to be evaluated according to the first evaluation keyword and a preset neural network model; Generate second evaluation data corresponding to the object to be evaluated according to the second evaluation keyword and a preset neural network model; The evaluation data includes the first evaluation data and the second evaluation data.

7. The data generation method according to claim 2, It is characterized in that Before generating evaluation data corresponding to the evaluation keyword according to the evaluation keyword and the preset neural network model, the method further includes: The preset neural network model is trained according to the evaluation keywords and the historical evaluation data, and the preset neural network model is a long short-term memory network model.

8. The data generation method according to claim 7, It is characterized in that The training of the preset neural network model according to the evaluation keywords includes: The evaluation keywords and each word in the word set are vectorized, and the vectorized data matrix is ​​used as training data input to train the long short-term memory network model.

9. A data generating device, It is characterized in that include: An acquisition module is used to obtain historical evaluation data of the object to be evaluated; An extraction module, used for extracting evaluation keywords of the historical evaluation data, wherein the evaluation keywords are used for characterizing the physical properties of the object to be evaluated; A processing module, used to generate evaluation data corresponding to the evaluation keywords according to the evaluation keywords and a preset neural network model; The evaluation data is the core evaluation data of the object to be evaluated corresponding to each of the evaluation keywords; A display module, configured to display the evaluation data at a first preset position in the evaluation area of ​​the object to be evaluated; The acquisition module is further used to acquire a first evaluation instruction input for first evaluation data, wherein the evaluation data includes the first evaluation data; The display module is further used to display a first sub-evaluation corresponding to the first evaluation instruction at a second preset position in the evaluation area of ​​the object to be evaluated, the first sub-evaluation being used to evaluate the first evaluation data, and the second preset position being adjacent to the first preset position in the evaluation area.

10. An electronic device, It is characterized in that include: processor; as well as, A memory, configured to store executable instructions of the processor; The processor is configured to perform the data generation method according to any one of claims 1 to 8 by executing the executable instructions.

11. A storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the data generating method according to any one of claims 1 to 8 is implemented.

Images