Question and answer text evaluation method, system and application

Abstract

The invention belongs to the technical field of text evaluation, and particularly relates to a question and answer text evaluation method and system and application. The method comprises the following steps: firstly, matching questions of an investor with MDamp of a corresponding year through a sliding window; a, extracting related texts; secondly, relevant information in the text is concerned through a variational auto-encoder and a bidirectional attention mechanism, and questions and answers and MDamp are mined; according to the method, the semantic difference between the question text and the answer text is obtained, the bidirectional attention weighted representation and the difference feature are spliced to obtain the context representation, and the semantic difference between the question text and the answer text is evaluated according to the context representation, so that the key information related to the share price volatility is accurately identified. And on the basis, quantitative analysis is performed on the mined semantic difference, so that accurate prediction of the share price volatility is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of text evaluation, and specifically relates to a question-and-answer text evaluation method, system and application. Background Art

[0002] At present, the prediction of stock price volatility mainly relies on constructing prediction features from structured data and then building a prediction model. These data include macroeconomic indicators, financial indicators, and stock market trading data. However, with the development of the new generation of information technology, scholars have begun to pay attention to the role of emotional information contained in text data in stock volatility prediction, such as conference call texts, social media texts, and earnings conference Q&A texts. In particular, earnings conference Q&A texts are considered to have potential in predicting stock price fluctuations because they reflect investors' focus issues and have the characteristics of instant two-way interaction.

[0003] Although existing technical means have begun to explore the effectiveness of text data in predicting stock price volatility, there are still some challenges. First, the non-standard, emotional, and ambiguous information in the spoken answer text makes the text analysis method face a lot of noise, making it difficult to accurately mine the semantic differences between questions and answers. Secondly, managers may selectively reveal some information and avoid negative details, making it difficult to accurately verify the authenticity and completeness of the information provided by relying solely on the answer text. In addition, when combining management analysis and discussion content to mine "irrelevant answers", there are also difficulties such as large amounts of information in long texts, complex content, and strong subjectivity. Therefore, how to effectively mine semantic differences from question and answer texts and predict stock price volatility is still a problem that current technology needs to solve. Summary of the invention

[0004] The purpose of the present invention is to provide a question-answer text evaluation method, system and application to solve the problems raised in the background technology.

[0005] The present invention achieves the above-mentioned purpose through the following technical solutions: In a first aspect, the present invention proposes a method for evaluating a question-answer text, comprising: S1. Obtain a topic text, a question text, and a reply text in a set question-and-answer scenario to form a question-and-answer dataset, wherein the question text is determined based on the topic text; S2, performing similarity matching between the question text and the subject text based on a preset text matching strategy, generating a continuous sentence sequence with the highest similarity as the matching text; S3. Preprocess the matching text and the question-answering dataset using the BERT model to generate a text vector; S4, inputting the text vector into a variational autoencoder for decoding and encoding to obtain a denoised text vector; S5, inputting the denoised text vector into a bidirectional long short-term memory network, performing forward and backward processing, and obtaining a hidden state sequence; S6. Calculate the bidirectional attention weighted representation and difference features of the hidden state sequence using a bidirectional attention mechanism; S7. Concatenate the bidirectional attention weighted representation and the difference feature to obtain a context representation, and evaluate the semantic difference between the question and answer text based on the context representation.

[0006] Furthermore, the set question-and-answer scenario includes a performance briefing of a listed company, the subject text includes a management discussion and analysis text, and the reply text includes a reply text of the highest-level management at the performance briefing of the listed company.

[0007] Further, the S2 includes: S201, deleting non-language symbols and numbers in the subject text; S202, using a sliding window-based text matching method to slide a fixed-size window on the subject text to calculate the similarity between the question text and the subject text; S203: Match the continuous sentence sequence with the highest similarity to the question text and use it as the matching text.

[0008] Furthermore, the preprocessing in S3 includes: performing word segmentation, removing stop words, and filtering low-frequency words on the matching text and the question-answer data set, respectively, to form a text vector containing contextual semantic information.

[0009] Further, the S4 includes: S401: Input the matching text and the text vector of the question-answering dataset into the encoder, perform linear transformation of the fully connected layer and apply the ReLU activation function to calculate the mean of the latent space. μ and log variance log(σ 2 ) , the latent variable z is obtained by sampling, ,in, is random noise sampled from a standard normal distribution; S402, obtaining the denoised text vector based on a fully connected layer and an activation function : ; ; in, represents the output of the third hidden layer; represents the weight matrix in the third hidden layer; represents the bias vector of the third hidden layer; Represents the weight matrix in the fourth hidden layer, specifically connecting the hidden layer and text vector The weight of Represents the bias vector of the fourth hidden layer.

[0010] Furthermore, S5 includes: converting the denoised text vector into a deep semantic representation through a bidirectional long short-term memory network, obtaining its context information for each vocabulary unit, forming a contextualized representation, and obtaining a hidden state sequence , , P , where A corresponds to the question text, Q corresponds to the answer text, and P corresponds to the matching text.

[0011] Further, the S6 includes: S601, calculation sequence arrive ,sequence and ,sequence arrive and sequence A arrive The dot product of the transpose of , we get the attention weight matrix; using Normalize the attention weight matrix so that the sum of the weights of each row is 1; S602, the obtained normalized weight matrix and the sequence Multiply them together and you get arrive The attention weighted representation of ,get arrive The attention weighted representation of , arrive The attention weighted representation of arrive The attention weighted representation of , as follows: ; ; ; ; S603, from sequence Subtract the weighted Q(A) from the original to get the difference feature , as follows: ; Further, the S7 includes: will sequence , Concatenate to get text representation ; Representing text Input to a fully connected layer, , get the corresponding similarity probability between texts and the probability of difference , as follows: ; .

[0012] In a second aspect, the present invention proposes a question-answer text evaluation system, which is applied to perform any of the above evaluation methods, and the system comprises: The data acquisition module is used to obtain the discussion and analysis texts of the management of listed companies, as well as the question texts and answer texts in the question and answer scenarios set by listed companies to form a question and answer data set; A text matching module, used to perform similarity matching between the question text and the subject text based on a preset text matching strategy, and generate a continuous sentence sequence with the highest similarity as the matching text; A preprocessing module, used to preprocess the matching text and the question-answering dataset using a BERT model to generate a text vector; A variational autoencoder module, used for inputting the text vector into a variational autoencoder, decoding and encoding, and obtaining a denoised text vector; A bidirectional long short-term memory network module is used to input the denoised text vector into the bidirectional long short-term memory network for forward and backward processing to obtain a hidden state sequence; A bidirectional attention mechanism module, used to calculate a bidirectional attention weighted representation and a difference feature of the hidden state sequence using a bidirectional attention mechanism; An evaluation module is used to concatenate the bidirectional attention weighted representation and the difference feature to obtain a context representation, and evaluate the semantic difference between the question and the answer text according to the context representation.

[0013] In a third aspect, the present invention proposes an application of a question-answer text evaluation method in stock price volatility prediction. The text representation G, the similarity probability between texts, and the difference probability between texts of a listed company's performance briefing are obtained by using any of the evaluation methods described above. The text representation G, the similarity probability between texts, and the difference probability between texts are concatenated to obtain a distributed text representation S: ; The identity characteristics of executives of listed companies After being merged into the text representation S, it is normalized and then input into the fully connected layer of the encoder. The stock price volatility is predicted based on the output of the fully connected layer, as shown in the following formula: ; in, represents the weight matrix, represents the bias vector, Represents the model output results.

[0014] The beneficial effects of the present invention are: The present invention combines the question-and-answer texts from the performance briefings of listed companies with the relevant texts in the MD&A through data matching, and further fills in the missing information and reduces noise through the variational autoencoder. The application of the bidirectional attention mechanism can more accurately understand the contextual relationship in the text, thereby accurately identifying the key information related to stock price volatility. On this basis, the excavated semantic differences are quantitatively analyzed to achieve accurate prediction of stock price volatility. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Schematic diagram of the process of the evaluation method of the present invention; Figure 2 Schematic diagram of the process of step S2 in the present invention; Figure 3 Schematic diagram of the process of step S4 in the present invention; Figure 4 Schematic diagram of the process of step S6 in the present invention; Figure 5 This is a stock price volatility prediction framework diagram based on the question and answer text of a listed company's performance briefing proposed in Example 3 of the present invention. DETAILED DESCRIPTION

[0016] The present application is further described in detail below in conjunction with the accompanying drawings. It is necessary to point out here that the following specific implementation methods are only used to further illustrate the present application and cannot be understood as limiting the scope of protection of the present application. Technical personnel in this field can make some non-essential improvements and adjustments to the present application based on the above application content.

[0017] Example 1 like Figure 1-4 As shown, this embodiment proposes a question-answer text evaluation method, comprising the following steps: S1. Obtain a topic text, a question text, and a reply text in a set question-and-answer scenario to form a question-and-answer dataset, wherein the question text is determined based on the topic text; As an example, the management discussion and analysis text (MD&A) of a listed company, as well as the question text and answer text in a question-and-answer scenario set by the listed company are obtained to form a question-and-answer dataset; the question-and-answer scenario is set to include a performance briefing of a listed company, and the answer text is the answer text of the highest-level management at the performance briefing of the listed company; S2, performing similarity matching between the question text and the subject text based on a preset text matching strategy, generating a continuous sentence sequence with the highest similarity as the matching text; S3. Preprocess the matching text and the question-answering dataset using the BERT model to generate a text vector; S4, inputting the text vector into a variational autoencoder for decoding and encoding to obtain a denoised text vector; S5, inputting the denoised text vector into a bidirectional long short-term memory network, performing forward and backward processing, and obtaining a hidden state sequence; S6. Calculate the bidirectional attention weighted representation and difference features of the hidden state sequence using a bidirectional attention mechanism; S7. Concatenate the bidirectional attention weighted representation and the difference feature to obtain a context representation, and evaluate the semantic difference between the question and answer text based on the context representation.

[0018] It should be noted that the Management Discussion and Analysis (MD&A) of listed companies is an important part of the annual report of listed companies, requiring management to further explain and analyze the important historical information in the company's current financial statements and notes, and explain the operating plan for the next year and the opportunities, challenges and various risks facing the company's future development from the perspective of the company's management.

[0019] According to the above steps S1-S7, the above evaluation method is mainly divided into three parts during its specific implementation, including data matching, data preprocessing and "irrelevant answer" scoring.

[0020] (1) In the data matching stage, this embodiment proposes a text matching method based on a sliding window, which matches the paragraph (matching text) that best matches the investor's question by calculating the similarity between the investor's question and the MD&A at the earnings conference. Preferably, S2 includes: S201. Delete the non-verbal symbols and numbers in the management discussion and analysis text; delete the "detailed description data" in MD&A. This type of detailed description data in MD&A may contain many non-verbal symbols, numbers, etc. These elements may have different distribution characteristics in the vector space, resulting in vector sparseness, affecting subsequent calculations and matching.

[0021] S202, using a sliding window-based text matching method to slide a fixed-size window on the management discussion and analysis text to calculate the similarity between the question text and the management discussion and analysis text; S203: Match the continuous sentence sequence with the highest similarity to the question text and use it as the matching text.

[0022] As an example, first, set the text length, that is, the window size , and segment the MD&A text and Q&A text into sentence units, i.e. and , and then calculate the similarity between each pair of sentences Next, a fixed-size window is slid across the MD&A text and the sum of the sentence similarities within each window is calculated, as shown in the following formula: ; Find the sequence of consecutive sentences with the highest similarity. This process is carried out within the text length limit, and the text in the final selected window is the part that best matches the question and answer text.

[0023] (2) In the data preprocessing stage, the executives are first classified, and then the text is segmented, stop words and low-frequency words are removed, and the text is converted into an embedding vector using the BERT model. Preferably, the preprocessing in S3 includes: segmenting the matching text and the question-answering dataset, removing stop words, and filtering low-frequency words, respectively, to form a text vector containing contextual semantic information.

[0024] (3) Using a non-answer evaluation method based on VAE (variational autoencoder) and bidirectional attention mechanism to mine the relationship between texts and calculate the degree of non-answer. Preferably, S4 includes: S401: Input the matching text and the text vector of the question-answering dataset into the encoder, perform linear transformation of the fully connected layer and apply the ReLU activation function to calculate the mean of the latent space. μ and log variance log(σ 2 ) , the latent variable z is obtained by sampling, ,in, is random noise sampled from a standard normal distribution; S402: Obtain the denoised text vector based on the fully connected layer and the activation function : ; ; in, represents the output of the third hidden layer, usually representing an intermediate layer in the network; Represents the weight matrix in the third hidden layer. In a neural network, weights are parameters in the learning process that are used to transform input data to the next layer. Represents the bias vector of the third hidden layer. The bias is another important learning parameter in neural networks. It is added to the product of the weight and the input to provide an offset for each neuron. Represents the weight matrix in the fourth hidden layer, specifically connecting the hidden layer and text vector The weight of Represents the bias vector of the fourth hidden layer.

[0025] During the training process, VAE optimizes two objectives at the same time: one is the reconstruction loss, which is used to measure the difference between the decoder output data and the original data. In this paper, it is measured by cross entropy, as shown below: ; in, is the reconstruction error, N is the number of data samples, D is the size of the vocabulary, The original data is The sample The value of the feature, is the decoder's reconstruction of The sample The other is the regularization term, which is measured by KL divergence in this embodiment to ensure that the distribution of the latent space is close to the prior distribution, as shown in the following formula: ; in, is the KL divergence, is the approximate posterior distribution of the encoder output, is a latent variable The prior distribution of is usually assumed to be a standard normal distribution. is the dimension of the latent space. is a latent variable The mean of the j-th dimension of . is a latent variable The variance of the j-th dimension.

[0026] By continuously adjusting the parameters of the encoder and decoder to minimize the above loss function, VAE can learn an effective representation of text data and capture the structure and semantic information of the text in the latent space. In order to further explore the semantic differences between texts, the two denoised text vectors are forward and backward processed by BiLSTM on the input sequence to capture the dependencies and information in the sequence.

[0027] Preferably, S5 includes: converting the denoised text vector into a deep semantic representation through a bidirectional long short-term memory network, obtaining its context information for each vocabulary unit, forming a contextualized representation, and obtaining a hidden state sequence , , P , where A corresponds to the question text, Q corresponds to the answer text, and P corresponds to the matching text.

[0028] The bidirectional attention mechanism can consider both investors’ questions and executives’ answers at the same time, which can capture the subtle semantic differences between the two. By building the relationship between questions and answers in the attention mechanism, it is possible to more clearly understand which question points are not fully discussed or completely ignored in the executives’ answers, thereby identifying the missing information.

[0029] Preferably, S6 includes: S601, calculation sequence arrive ,sequence and ,sequence arrive and sequence A arrive The dot product of the transpose of , we get the attention weight matrix; using Normalize the attention weight matrix so that the sum of the weights of each row is 1; S602, the obtained normalized weight matrix and the sequence Multiply them together and you get arrive The attention weighted representation of ,get arrive The attention weighted representation of , arrive The attention weighted representation of arrive The attention weighted representation of , as follows: ; ; ; ; S603, from sequence Subtract the weighted Q(A) from the original to get the difference feature , as follows: ; Further preferably, the S7 includes: will sequence , Concatenate to get text representation ; ; Representing text Input to a fully connected layer, , get the corresponding similarity probability between texts and the probability of difference , as follows: ; .

[0030] According to the above embodiment of the present invention, this embodiment proposes a method for evaluating question and answer texts based on performance briefings of listed companies. The method first extracts relevant texts by matching investors' questions and corresponding MD&A through a sliding window; secondly, the variational autoencoder and the bidirectional attention mechanism focus on the relevant information in the text, and mine the irrelevant answers between questions and answers, and between answers and MD&A related texts, and evaluates the question and answer texts based on similarity probability and difference probability. Example

[0031] This embodiment proposes a question-answer text evaluation system, which is applied to perform the above evaluation method. The system includes: A data acquisition module, used to acquire a subject text, a question text and a reply text in a set question-answering scenario to form a question-answering data set, wherein the question text is determined based on the subject text; A text matching module, used to perform similarity matching between the question text and the subject text based on a preset text matching strategy, and generate a continuous sentence sequence with the highest similarity as the matching text; A preprocessing module, used to preprocess the matching text and the question-answering dataset using a BERT model to generate a text vector; A variational autoencoder module, used for inputting the text vector into a variational autoencoder, decoding and encoding, and obtaining a denoised text vector; A bidirectional long short-term memory network module is used to input the denoised text vector into the bidirectional long short-term memory network for forward and backward processing to obtain a hidden state sequence; A bidirectional attention mechanism module, used to calculate a bidirectional attention weighted representation and a difference feature of the hidden state sequence using a bidirectional attention mechanism; An evaluation module is used to concatenate the bidirectional attention weighted representation and the difference feature to obtain a context representation, and evaluate the semantic difference between the question and the answer text according to the context representation.

[0032] It should be noted here that each module in the above evaluation system corresponds to steps S1 to S7 in the implementation of the above evaluation method, and the instances and application scenarios implemented by multiple modules and corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment 1. Example

[0033] Combination Figure 5 This embodiment proposes an application of a question-answer text evaluation method in stock price volatility prediction. The evaluation method described in Example 1 is used to obtain the text representation G, the similarity probability between texts, and the difference probability between texts of the performance briefing of listed companies. The text representation G, the similarity probability between texts, and the difference probability between texts are concatenated to obtain a distributed text representation S: ; The identity characteristics of executives of listed companies After being merged into the text representation S, it is normalized and then input into the fully connected layer of the encoder. The stock price volatility is predicted based on the output of the fully connected layer, as shown in the following formula: ; in, represents the weight matrix, represents the bias vector, Represents the model output results.

[0034] More specifically, based on the nonlinear characteristics of the stock market volatility, which presents a peak and a thick tail, this embodiment uses the natural logarithm form to quantify the stock price volatility. , as shown below: ; in Indicates time, Indicates The rate of change of the day (corresponding to the output Y of the above fully connected layer), It represents the average rate of change during this period, using the volatility of the week after the performance description occurs ( ).

[0035] According to the above-mentioned embodiments of the present invention, this embodiment proposes a question-answer text evaluation method based on Embodiment 1 to predict stock price volatility, and realizes the prediction of stock price volatility using question-answer text. Specifically, based on variational autoencoder and bidirectional attention mechanism, it extracts irrelevant information and calculates difference probability, and has strong prediction ability.

[0036] In another embodiment provided by the present invention, a computer-readable storage medium is also provided, in which a computer program is stored. When the computer program is executed by a processor, the steps of the evaluation method of the question and answer text in the above-mentioned embodiment 1 are implemented.

[0037] In another embodiment of the present invention, a computer program product including instructions is provided. When the computer program product is run on a computer, the computer executes the method for evaluating the question and answer text in the above-mentioned embodiment 1.

[0038] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated.

[0039] The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0040] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0041] In addition, each functional module in each embodiment of the present application 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.

[0042] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features can be replaced by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for evaluating a question-answer text, characterized in that: include: S1. Obtain a topic text, a question text, and a reply text in a set question-and-answer scenario to form a question-and-answer dataset, wherein the question text is determined based on the topic text; S2, performing similarity matching between the question text and the subject text based on a preset text matching strategy, generating a continuous sentence sequence with the highest similarity as the matching text; S3. Preprocess the matching text and the question-answering dataset using the BERT model to generate a text vector; S4, inputting the text vector into a variational autoencoder for decoding and encoding to obtain a denoised text vector; S5, inputting the denoised text vector into a bidirectional long short-term memory network, performing forward and backward processing, and obtaining a hidden state sequence; S6. Calculate the bidirectional attention weighted representation and difference features of the hidden state sequence using a bidirectional attention mechanism; S7. Concatenate the bidirectional attention weighted representation and the difference feature to obtain a context representation, and evaluate the semantic difference between the question and answer text based on the context representation.

2. The method for evaluating a question-answer text according to claim 1, characterized in that: The set question-and-answer scenario includes a performance briefing of a listed company, the subject text includes a management discussion and analysis text, and the reply text includes a reply text of the highest-level management at the performance briefing of the listed company.

3. The method for evaluating a question-answer text according to claim 2, characterized in that: The S2 includes: S201, deleting non-language symbols and numbers in the subject text; S202, using a sliding window-based text matching method to slide a fixed-size window on the subject text to calculate the similarity between the question text and the subject text; S203: Match the continuous sentence sequence with the highest similarity to the question text and use it as the matching text.

4. A question-answer text evaluation method according to claim 2, characterized in that: The preprocessing in S3 includes: performing word segmentation, removing stop words, and filtering low-frequency words on the matching text and the question-answer data set, respectively, to form a text vector containing contextual semantic information.

5. A question-answer text evaluation method according to claim 3, characterized in that: The S4 includes: S401: Input the matching text and the text vector of the question-answering dataset into the encoder, perform linear transformation of the fully connected layer and apply the ReLU activation function to calculate the mean of the latent space. μ and log variance log(σ 2 ) , the latent variable z is obtained by sampling, ,in, is random noise sampled from a standard normal distribution; S402: Obtain the denoised text vector based on the fully connected layer and the activation function : ； ； in, represents the output of the third hidden layer; represents the weight matrix in the third hidden layer; represents the bias vector of the third hidden layer; Represents the weight matrix in the fourth hidden layer, specifically connecting the hidden layer and text vector The weight of Represents the bias vector of the fourth hidden layer.

6. A question-answer text evaluation method according to claim 5, characterized in that: S5 includes: converting the denoised text vector into a deep semantic representation through a bidirectional long short-term memory network, obtaining context information for each vocabulary unit, forming a contextual representation, and obtaining a hidden state sequence , , P , where A corresponds to the question text, Q corresponds to the answer text, and P corresponds to the matching text.

7. A question-answer text evaluation method according to claim 6, characterized in that: The S6 includes: S601, calculation sequence arrive ,sequence arrive ,sequence arrive and sequence A arrive The dot product of the transpose of , we get the attention weight matrix; using Normalize the attention weight matrix so that the sum of the weights of each row is 1; S602, the obtained normalized weight matrix and the sequence Multiply them together and you get arrive The attention weighted representation of ,get arrive The attention weighted representation of , arrive The attention weighted representation of arrive The attention weighted representation of , as follows: ； ； ； ； S603, from sequence Subtract the weighted Q(A) from the original to get the difference feature , as follows: 。 8. A question-answer text evaluation method according to claim 7, characterized in that: The S7 includes: will sequence , Concatenate to get text representation ; ； Representing text Input to a fully connected layer, , get the corresponding similarity probability between texts and the probability of difference , as follows: ； 。 9. A question-answer text evaluation system, characterized in that: Applied to executing the evaluation method according to any one of claims 1 to 8, the system comprises: The data acquisition module is used to obtain the discussion and analysis texts of the management of listed companies, as well as the question texts and answer texts in the question and answer scenarios set by listed companies to form a question and answer data set; A text matching module, used to perform similarity matching between the question text and the subject text based on a preset text matching strategy, and generate a continuous sentence sequence with the highest similarity as the matching text; A preprocessing module, used to preprocess the matching text and the question-answering dataset using a BERT model to generate a text vector; A variational autoencoder module, used for inputting the text vector into a variational autoencoder, decoding and encoding, and obtaining a denoised text vector; A bidirectional long short-term memory network module is used to input the denoised text vector into the bidirectional long short-term memory network for forward and backward processing to obtain a hidden state sequence; A bidirectional attention mechanism module, used to calculate a bidirectional attention weighted representation and a difference feature of the hidden state sequence using a bidirectional attention mechanism; An evaluation module is used to concatenate the bidirectional attention weighted representation and the difference feature to obtain a context representation, and evaluate the semantic difference between the question and the answer text based on the context representation.

10. An application of a question-answer text evaluation method in predicting stock price volatility, characterized in that: The evaluation method according to any one of claims 1 to 8 is used to obtain the text representation G, the similarity probability between texts, and the difference probability between texts of the performance briefing of the listed company, and the text representation G, the similarity probability between texts, and the difference probability between texts are concatenated to obtain the distributed text representation S: ； The identity characteristics of executives of listed companies After being merged into the text representation S, it is normalized and then input into the fully connected layer of the encoder. The stock price volatility is predicted based on the output of the fully connected layer, as shown in the following formula: ; in, represents the weight matrix, represents the bias vector, Represents the model output results.

Images