Methods, devices and computer equipment for processing script documents

Abstract

This invention belongs to the field of document processing and provides a method for processing employee script documents. The method includes: obtaining the weights of keywords in an employee's script document; determining keyword scores in the employee's script document based on the keyword weights; classifying the employee's script document to obtain a semantic category score; and determining the employee's competency indicators based on the keyword scores and the semantic category scores. This invention enables the evaluation of script documents and the analysis of employee competency based on the evaluation results. This method allows for the rapid and accurate analysis of employee competency indicators, facilitating the development of subsequent business strategies based on these indicators.

Description

Technical Field

[0001] This invention belongs to the field of document processing, and more specifically relates to a method, apparatus and computer equipment for processing script documents. Background Technology

[0002] With the rapid development of the internet, various business scenarios have corresponding sales scripts. In these scenarios, sales scripts are crucial to the direction of business operations; good scripts can expedite process execution. For example, in telemarketing, sales scripts significantly impact the success rate of closing a deal; effective scripts increase the probability of success, while poor scripts may decrease it. Currently, there is a pressing need for an evaluation method to assess sales scripts and then analyze employee capabilities based on the evaluation results.

[0003] Therefore, it is necessary to provide a method for processing script documents to solve the above problems. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] The present invention aims to solve the problem that there is currently no evaluation method in the related technology to evaluate dialogue and analyze the employee's ability based on the evaluation results.

[0006] (II) Technical Solution

[0007] To address the aforementioned technical problems, one aspect of the present invention proposes a method for processing employee script documents, comprising: obtaining the weights of keywords in an employee script document; determining keyword scores in the employee script document based on the keyword weights; classifying the employee script document to obtain a semantic category score; and determining the employee's competency indicators based on the keyword scores and the semantic category scores of the employee script document.

[0008] According to a preferred embodiment of the present invention, obtaining the weight of a keyword based on the keywords in the employee script document includes: extracting the keywords from the employee script document according to a keyword thesaurus; determining the weight of the keyword according to a mapping table between keywords and weights, wherein the weight of the keyword is the degree of difference of the keyword in multiple batches of historical script documents.

[0009] According to a preferred embodiment of the present invention, the method further includes: obtaining a word set based on the multiple batches of historical speech documents; determining the degree of difference of each word in the word set in different batches of historical speech documents based on the inverse document frequency and word frequency of each word in the word set in different batches of historical speech documents; and obtaining a keyword lexicon and a keyword-weight mapping table based on the degree of difference of each word in the word set in different batches of historical speech documents.

[0010] According to a preferred embodiment of the present invention, if the employee script document contains multiple keywords, determining the keyword score in the employee script document based on the weight of the keywords includes: calculating the sum of the weights of the keywords based on the weight of each keyword among the multiple keywords; and determining the keyword score in the employee script document based on the sum of the weights of the keywords.

[0011] According to a preferred embodiment of the present invention, classifying the employee script document to obtain a semantic category score for the employee script document includes: inputting the employee script document into a prediction model, obtaining semantic information in the employee script document through the prediction model, wherein the prediction model is a BERT model or a neural network model; and determining the semantic category score of the employee script document based on the semantic information of the employee script document.

[0012] According to a preferred embodiment of the present invention, determining an employee's competency indicators based on the keyword scores and semantic category scores in the employee's script document includes: concatenating the keyword scores and semantic category scores in the employee's script document to obtain a feature vector of the employee's script document; inputting the feature vector of the employee's script document into a logistic regression model to obtain a comprehensive score of the employee's script document through the logistic regression model; and matching the competency indicators corresponding to the employee based on the comprehensive score of the employee's script document.

[0013] According to a preferred embodiment of the present invention, obtaining the comprehensive score of the employee's script document through the logistic regression model includes: calculating the comprehensive score of the employee's script document using the logistic regression algorithm in the logistic regression model, wherein the logistic regression algorithm is:

[0014]

[0015] Where x is the feature vector of the employee's script document, and w and b are the parameters in the logistic regression model.

[0016] A second aspect of the present invention provides a speech document processing device, comprising: a keyword weight acquisition module, used to acquire the weight of the keywords in the employee speech document; a keyword score determination module, used to determine the keyword score in the employee speech document based on the keyword weight; a semantic category acquisition module, used to classify the employee speech document to obtain a semantic category score of the employee speech document; and a competency indicator determination module, used to determine the employee's competency indicators based on the keyword score and the semantic category score of the employee speech document.

[0017] A third aspect of the present invention provides a computer device, including a processor and a memory, wherein the memory is used to store a computer-executable program, and when the computer program is executed by the processor, the processor executes a speech document processing method as described in any of the preceding claims.

[0018] A fourth aspect of the present invention provides a computer program product storing a computer-executable program, wherein when the computer-executable program is executed, it implements a script document processing method as described in any of the preceding claims.

[0019] (III) Beneficial Effects

[0020] Compared to related technologies, this invention obtains the weights of keywords in employee script documents, determines keyword scores based on these weights, categorizes the employee script documents to obtain semantic category scores, and then determines the employee's competency indicators based on both the keyword scores and the semantic category scores. This invention enables the evaluation of script documents and the analysis of employee competency based on the evaluation results. This method allows for the rapid and accurate analysis of employee competency indicators, facilitating the development of subsequent business strategies based on these indicators. Attached Figure Description

[0021] Figure 1 A schematic diagram of an exemplary system architecture to which the technical solutions of embodiments of the present invention can be applied is shown;

[0022] Figure 2 This is a flowchart illustrating an example of a speech document processing method according to an embodiment of the present invention;

[0023] Figure 3 This is a flowchart illustrating another example of the speech document processing method according to an embodiment of the present invention;

[0024] Figure 4 This is a flowchart illustrating another example of the speech document processing method according to an embodiment of the present invention;

[0025] Figure 5This is a flowchart illustrating another example of the speech document processing method according to an embodiment of the present invention;

[0026] Figure 6 This is a flowchart illustrating another example of the speech document processing method according to an embodiment of the present invention;

[0027] Figure 7 This is a flowchart illustrating another example of the speech document processing method according to an embodiment of the present invention;

[0028] Figure 8 This is a schematic diagram of an example of a speech document processing device according to an embodiment of the present invention;

[0029] Figure 9 This is a schematic diagram of yet another example of the speech document processing device according to an embodiment of the present invention;

[0030] Figure 10 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention;

[0031] Figure 11 This is a schematic diagram of a computer program product according to an embodiment of the present invention. Detailed Implementation

[0032] In the description of specific embodiments, detailed descriptions of structures, performance, effects, or other features are provided to enable those skilled in the art to fully understand the embodiments. However, this does not preclude those skilled in the art from implementing the present invention with technical solutions that do not contain the aforementioned structures, performance, effects, or other features under specific circumstances.

[0033] The flowcharts in the accompanying drawings are merely illustrative examples and do not imply that the solution of this invention must include all the content, operations, and steps shown in the flowcharts, nor do they imply that the execution must be performed in the order shown in the diagrams. For example, some operations / steps in the flowcharts can be decomposed, some operations / steps can be combined or partially combined, etc. Without departing from the inventive spirit of this invention, the execution order shown in the flowcharts can be changed according to the actual situation.

[0034] The box in the attached diagram Figure 1 Generally, these refer to functional entities, and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processing unit devices and / or microcontroller devices.

[0035] The same reference numerals in the accompanying drawings denote the same or similar elements, components, or parts, and therefore, repeated descriptions of the same or similar elements, components, or parts may be omitted below. It should also be understood that although terms such as first, second, third, etc., indicating numbers may be used herein to describe various devices, elements, components, or parts, these devices, elements, components, or parts should not be limited by these terms. That is, these terms are only used to distinguish one from another. For example, a first device may also be referred to as a second device, without departing from the essential technical solution of the invention. Furthermore, the terms "and / or" and "and / or" refer to all combinations including any one or more of the listed items.

[0036] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0037] Figure 1 A schematic diagram of an exemplary system architecture to which the technical solutions of embodiments of the present invention can be applied is shown.

[0038] like Figure 1 As shown, system architecture 100 may include one or more of user terminals 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between user terminals 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0039] It should be understood that Figure 1 The number of user terminals, networks, and servers shown is merely illustrative. Depending on implementation needs, there can be any number of user terminals, networks, and servers. For example, server 105 could be a server cluster composed of multiple servers.

[0040] Users can use user terminals 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. User terminals 101, 102, and 103 can be various electronic devices with displays, including but not limited to smartphones, tablets, laptops, and desktop computers, etc.

[0041] Server 105 can be a server that provides various services. For example, server 105 can obtain the weight of keywords in the employee's script document; determine the keyword score in the employee's script document based on the keyword weight; classify the employee's script document to obtain the semantic category score; and determine the employee's competency indicators based on the keyword score and the semantic category score. This method can quickly and accurately analyze the employee's competency indicators, which facilitates the formulation of subsequent business strategies based on the employee's competency indicators.

[0042] In some embodiments, the script document processing method provided in this invention is generally executed by server 105, and correspondingly, the script document processing device is generally located in server 105. In other embodiments, some terminals may have functions similar to those of the server to execute this method. Therefore, the script document processing method provided in this invention is not limited to execution on the server side.

[0043] Figure 2 This is a flowchart illustrating an example of a speech document processing method according to an embodiment of the present invention.

[0044] like Figure 2 As shown, the script document processing method includes steps S210 to S240.

[0045] In step S210, the weight of the keywords is obtained based on the keywords in the employee script document.

[0046] In step S220, the keyword scores in the employee script document are determined based on the weights of the keywords.

[0047] In step S230, the employee script document is classified to obtain a semantic category score for the employee script document.

[0048] In step S240, the employee's competency indicators are determined based on the keyword scores and semantic category scores in the employee's script document.

[0049] This method enables the evaluation of script documents and the analysis of employee capabilities based on the evaluation results. For example, it obtains the weights of keywords in an employee's script document, determines the keyword scores based on these weights, categorizes the script documents to obtain semantic category scores, and then determines the employee's capability indicators based on both the keyword scores and the semantic category scores. This method allows for the rapid and accurate analysis of employee capability indicators, facilitating the development of subsequent business strategies based on these indicators.

[0050] In some embodiments of the present invention, the aforementioned employee script document can be the text of dialogue generated by the employee during the course of work. For example, if the employee is an online customer service representative, the script document can be the dialogue text between the online customer service representative and the user. As another example, if the employee is an online teacher, the script document can be the dialogue text between the online teacher and the parent or student. Yet another example, if the employee is an online doctor, the script document can be the dialogue text between the online doctor and the patient. In this embodiment, the aforementioned script document can be the text of dialogue from an employee in any field.

[0051] In some embodiments of the present invention, before extracting keywords from employee script documents, a keyword lexicon can be obtained based on multiple batches of historical script documents. For example, a word set is obtained based on multiple batches of historical script documents. The degree of difference of each word in the word set in different batches of historical script documents is determined based on the inverse document frequency and word frequency of each word in the word set in different batches of historical script documents. Then, a keyword lexicon is obtained based on the degree of difference of each word in the word set in different batches of historical script documents.

[0052] In some embodiments of the present invention, a mapping table between keywords and weights can also be obtained based on the degree of difference of each word in the word set in different batches of historical speech documents.

[0053] Based on the aforementioned embodiments, each word in the employee's script document is traversed, and words that are identical to keywords in the aforementioned thesaurus are selected from the employee's script document. These selected words are then used as keywords in the employee's script document. Next, based on the keywords in the employee's script document, the corresponding weights are retrieved from the aforementioned keyword-weight mapping table. In this way, the weight of each keyword in the employee's script document can be obtained. In this embodiment, the weight of each keyword in the employee's script document can be the degree of difference of the keyword in different historical script documents. For example, if the degree of difference is large, it means that the meaning expressed by the keyword differs greatly in different historical script documents. Conversely, if the degree of difference is small, it means that the meaning expressed by the keyword differs little in different historical script documents.

[0054] In some embodiments of the present invention, keyword scores in employee script documents are determined based on keyword weights. For example, the weights of each keyword in the employee script document are summed, and the calculated sum of keyword weights is used as the keyword score for the employee script document. In this embodiment, a higher keyword score in the employee script document indicates that the keywords selected through the above method better express the core content of the employee script document.

[0055] In some embodiments of the present invention, employee script documents are classified to obtain semantic category scores. For example, by using a pre-trained prediction model to classify employee script documents, the probabilities of multiple category labels can be obtained. In this embodiment, the probability of the positive category label among the probabilities of multiple category labels is used as the semantic category score of the employee script document. It should be noted that the positive category label can be set according to the specific business scenario. For example, if the business scenario is a sales-related business scenario, the positive category label can be the probability of closing a deal, and the negative category label can be the probability of not closing a deal.

[0056] The script document processing method described in this invention can analyze an employee's capabilities from two aspects of scoring, making the analysis results closer to the employee's true level. In this invention, the two aspects of scoring can be keyword scoring and semantic category scoring of the employee's script document. For example, based on the keyword scoring and semantic category scoring of the employee's script document, the employee's capability indicators can be quickly and accurately analyzed. These capability indicators can be set according to specific business scenarios. For example, one-star, two-star, three-star, etc. A higher star rating indicates stronger capabilities, and conversely, a lower star rating indicates weaker capabilities.

[0057] In some embodiments of the present invention, training strategies are formulated based on keyword scores and semantic category scores in employee script documents. This allows for targeted training of employees and recommendation of high-scoring scripts to other employees' devices, thereby helping to improve employees' overall capabilities.

[0058] Figure 3 This is a flowchart of another example of the speech document processing method according to an embodiment of the present invention.

[0059] like Figure 3 As shown, step S210 can specifically include steps S310 to S320.

[0060] In step S310, the keywords are extracted from the employee script document according to the keyword thesaurus.

[0061] In step S320, the weight of the keyword is determined according to the mapping table between keywords and weights. The weight of the keyword is the degree of difference of the keyword in multiple batches of historical script documents.

[0062] This method can extract keywords from employee script documents based on a keyword thesaurus. In this way, words that reflect the core content of the employee script document can be extracted. Then, according to the keyword-weight mapping table, the weight of the keyword is determined. This makes it easier to accurately score the keywords in the employee script document based on the weight of each keyword in the document.

[0063] In some embodiments of the present invention, the vocabulary of the aforementioned keywords is obtained based on the degree of difference of each word in the word set in different batches of historical script documents. For example, the degree of difference of each word in the word set in different batches of historical script documents is sorted, and the top K words are added to the vocabulary of the aforementioned keywords based on the sorting results. In this embodiment, the top K can be set according to the actual situation.

[0064] In some embodiments of the present invention, each word in the employee's script document is traversed, and words that are the same as the keywords in the aforementioned thesaurus are selected from the employee's script document. These selected words are then used as the keywords in the employee's script document. Based on the keywords in the employee's script document, the corresponding weights are queried from the aforementioned keyword-weight mapping table. In this way, the weight of each keyword in the employee's script document can be obtained.

[0065] Figure 4 This is a flowchart of another example of the speech document processing method according to an embodiment of the present invention.

[0066] like Figure 4 As shown, the above method also includes steps S410 to S430.

[0067] In step S410, a set of words is obtained based on the multiple batches of historical script documents.

[0068] In step S420, the degree of difference of each word in the word set in different batches of historical speech documents is determined based on the inverse document frequency of each word in the word set in different batches of historical speech documents and the word frequency of each word in the word set in different batches of historical speech documents.

[0069] In step S430, based on the degree of difference of each word in the word set in different batches of historical speech documents, the keyword lexicon and the mapping table between the keywords and weights are obtained.

[0070] This method can add words with greater differences to the keyword lexicon based on the degree of difference of each word in the word set across different batches of historical script documents. This allows for the accurate and rapid extraction of highly distinctive keywords from employee script documents based on the lexicon.

[0071] In some embodiments of the present invention, the aforementioned batches of historical sales script documents may include at least two types of historical sales script documents. For example, the batches of historical sales script documents may be completed sales historical sales script documents and uncompleted sales sales historical sales script documents. Word segmentation is performed on the completed sales historical sales script documents and the uncompleted sales sales historical sales script documents respectively to obtain words from the completed sales historical sales script documents and the uncompleted sales sales historical sales script documents. Based on the words from the completed sales historical sales script documents and the uncompleted sales sales historical sales script documents, the aforementioned word set is constructed.

[0072] In some embodiments of the present invention, before determining the degree of difference of each word in the above-mentioned word set in different batches of historical speech documents, the method further includes: calculating the inverse document frequency of each word in the above-mentioned word set in different batches of historical speech documents; and calculating the word frequency of each word in the word set in different batches of historical speech documents.

[0073] Based on the aforementioned embodiments, the inverse document frequency of each word in the above word set in different batches of documents is calculated using formula (1), as shown below:

[0074]

[0075] Wherein, IDF represents inverse document frequency, w represents a word in the above word set, and b represents a batch identifier for multiple batches of historical script documents, such as those that did not result in a sale. This represents the inverse document frequency of word w in the word set within the unformed speech document. In this embodiment, This represents the inverse document frequency of word w in the word set within the single-word document.

[0076] In some embodiments of the present invention, the word frequency of each word in the above word set in different batches of documents is calculated by formula (2), as shown below:

[0077]

[0078] Where TF represents word frequency, w represents words in the above word set, and b represents batch identifiers for multiple batches of historical script documents, such as those that did not result in a sale. This represents the word frequency of word w in the unformed speech document within the word set. In this embodiment, This indicates the word frequency of word w in the word set within the single-word script document.

[0079] In some embodiments of the present invention, the degree of difference of each word in the word set across different batches of historical script documents is determined based on the inverse document frequency and word frequency of each word in the word set across different batches of historical script documents. For example, the degree of difference of word w in a completed script document relative to word w in a non-completed script document is calculated based on the inverse document frequency and word frequency of each word in the word set across different batches of historical script documents. As another example, the degree of difference of word w in a non-completed script document relative to word w in a completed script document is calculated based on the inverse document frequency and word frequency of each word in the word set across different batches of historical script documents.

[0080] In some embodiments of the present invention, the difference between words w in a single-word script document and words w in a non-single-word script document is calculated based on the inverse document frequency of each word in the word set in different batches of historical script documents and the word frequency of each word in the word set in different batches of historical script documents. For example, the difference between words w in a single-word script document and words w in a non-single-word script document is calculated using formula (3), which is shown below:

[0081]

[0082] Where DP represents the degree of difference. This represents the degree of difference between word w in the sales script document and word w in the non-sales script document. The word frequency of word 'w' in the word set within the order-taking script document. This indicates the word frequency of word w in the word set within the unformed speech document. This represents the inverse document frequency of word w in the word set within the single-speech document. This represents the inverse document frequency of word w in the unformed speech document.

[0083] In some embodiments of the present invention, the difference between words w in unsuccessful sales script documents and words w in successful sales script documents is calculated based on the inverse document frequency of each word in the word set in different batches of historical sales script documents and the word frequency of each word in the word set in different batches of historical sales script documents. The difference between words w in unsuccessful sales script documents and words w in successful sales script documents is calculated using formula (4), which is shown below:

[0084]

[0085] Where DP represents the degree of difference. This indicates the degree of difference between word w in the unsuccessful sales script and word w in the successful sales script. The word frequency of word 'w' in the word set within the order-taking script document. This indicates the word frequency of word w in the word set within the unformed speech document. This represents the inverse document frequency of word w in the word set within the single-speech document. This represents the inverse document frequency of word w in the unformed speech document.

[0086] In some embodiments of the present invention, a keyword lexicon and a keyword-weight mapping table are obtained based on the degree of difference of each word in the word set in different batches of historical script documents. For example, the degree of difference of each word in the word set in different batches of historical script documents is sorted. For example, the different batches of historical script documents are sales script documents and non-sales script documents, and the word set contains five words, namely word u, word v, word w, word x, and word y. The five degree of difference are calculated by the above formula (3). Arrange in descending order Sort the data, and the sorting result is: Based on this ranking result, the top K words can be selected as differentiated keywords in different batches of historical script documents. For example, words w, x, and v can be used as differentiated keywords in different batches of historical script documents, and then a thesaurus of these keywords can be constructed. And based on words w, x, v, and... Construct a mapping table between the aforementioned keywords and their weights. In this embodiment, topK can be set according to actual circumstances.

[0087] Figure 5 This is a flowchart of another example of the speech document processing method according to an embodiment of the present invention.

[0088] like Figure 5 As shown, if the employee script document contains multiple keywords, the above step S220 may specifically include steps S510 to S520.

[0089] In step S510, the weights of the keywords are calculated based on the weight of each keyword among the plurality of keywords.

[0090] In step S520, the keyword scores in the employee script document are determined based on the sum of the weights of the keywords.

[0091] This method can determine the keyword scores in employee script documents based on the sum of keyword weights, resulting in more accurate keyword scores.

[0092] In some embodiments of the present invention, the weight corresponding to each keyword is queried from the keyword-weight mapping table among multiple keywords, and the weights of each keyword among the multiple keywords are summed to obtain the sum of the keyword weights. In this embodiment, the sum of keyword weights can be used as a keyword score in the employee's script document; a higher score indicates a stronger ability of the employee.

[0093] Figure 6 This is a flowchart of another example of the speech document processing method according to an embodiment of the present invention.

[0094] like Figure 6 As shown, step S230 may specifically include steps S610 to S620.

[0095] In step S610, the employee script document is input into the prediction model, and the semantic information in the employee script document is obtained through the prediction model.

[0096] In step S620, the semantic category score of the employee script document is determined based on the semantic information of the employee script document.

[0097] This method can determine the semantic category score of employee script documents based on the semantic information of the script documents, and the semantic category score obtained in this way is more accurate.

[0098] In some embodiments of the present invention, the prediction model is a BERT model or a neural network model. Before obtaining semantic information from employee script documents, a BERT model or a neural network model can be trained using historical script documents. In this embodiment, training the model may include the following steps:

[0099] i. Use the aforementioned historical script documents as training samples.

[0100] ii. Set the model's labels based on the specific business type of the historical sales scripts mentioned above. For example, use whether the training sample resulted in a sale as the category label, such as "sold" and "not sold".

[0101] iii. Input the training sample and the set labels into the BERT model or neural network model, start training until the BERT model or neural network model converges, and then stop training.

[0102] In some embodiments of the present invention, semantic information in employee script documents is obtained through a prediction model. For example, the employee script documents are processed by an algorithm in the prediction model, and the processing result is used as the semantic information in the employee script documents.

[0103] In some embodiments of the present invention, the semantic category score of the employee's script document is determined based on the semantic information of the script document. For example, the probability of the category label to which the semantic information of the employee's script document belongs is calculated using the algorithm in the prediction model described above, and the semantic category score of the employee's script document is determined based on the probability of the category label to which the semantic information of the employee's script document belongs.

[0104] Based on the above embodiments, the preset model's category labels are "successful sale" and "failed sale." The semantic category score of the employee's script document is determined based on the probability of the semantic information of that document belonging to that category label. For example, the success rate of a sale output by a BERT model or a neural network model can be used as the semantic category score of the employee's script document. A higher semantic category score indicates a stronger employee's ability.

[0105] Figure 7 This is a flowchart of another example of the speech document processing method according to an embodiment of the present invention.

[0106] like Figure 7 As shown, step S240 may specifically include steps S710 to S730.

[0107] In step S710, the keyword scores and semantic category scores of the employee script document are concatenated to obtain the feature vector of the employee script document.

[0108] In step S720, the feature vector of the employee's script document is input into the logistic regression model, and the comprehensive score of the employee's script document is obtained through the logistic regression model.

[0109] In step S730, based on the comprehensive score of the employee's script document, the corresponding competency indicators are matched with the employee.

[0110] This method provides a comprehensive analysis of employee capabilities from two perspectives, resulting in a more accurate assessment of their true abilities. For example, keyword scores and semantic category scores from an employee's script are used as features input to a logistic regression model. This model analyzes the feature vector generated from these scores to obtain a comprehensive score for the script. Then, based on this comprehensive score, corresponding capability indicators are matched to the employee, leading to more accurate determination of their capabilities.

[0111] In some embodiments of the present invention, before obtaining the comprehensive score of employee sales scripts through a logistic regression model, the logistic regression model is first trained using historical sales scripts. For example, taking the training of a logistic regression model for closing deals as an example, the training process may include the following steps:

[0112] i. Use the aforementioned historical script documents as training samples.

[0113] ii. through Figure 5 The method described in the embodiment obtains keyword scores from training samples.

[0114] iii. Through Figure 6 The method described in the embodiment obtains the semantic category score of the training samples.

[0115] iv. Set the model's labels based on the specific business type of the historical sales scripts mentioned above. For example, use whether the training sample resulted in a sale as the category label, such as "sold" and "not sold".

[0116] v. Use the keyword scores and semantic category scores of the training samples as features of the training samples, and input these features and category labels into the logistic regression model for training until the logistic regression model converges and the parameters w and b in the logistic regression model fluctuate within a certain range, then stop training.

[0117] vi. Use the probability values ​​output by the logistic regression model as the overall score for the training samples.

[0118] In some embodiments of the present invention, a competency indicator corresponding to the employee is matched based on the overall score of the employee's script document. For example, the competency indicator corresponding to the overall score is queried from a mapping table between the overall score and competency indicators based on the overall score of the employee's script document. The retrieved competency indicator is then used as the competency indicator corresponding to the employee. In this embodiment, the competency indicator can be set according to specific business scenarios, such as excellent, good, average, etc.

[0119] In some embodiments of the present invention, obtaining the comprehensive score of the employee's script document through a logistic regression model includes: calculating the comprehensive score of the employee's script document using a logistic regression algorithm in the logistic regression model, wherein the logistic regression algorithm is:

[0120]

[0121] Where x is the feature vector of the employee's script document, and w and b are the parameters in the logistic regression model.

[0122] Figure 8 This is a schematic diagram of an example of a speech document processing apparatus according to an embodiment of the present invention.

[0123] like Figure 8 As shown, the script document processing device 800 includes a keyword weight acquisition module 810, a keyword score determination module 820, a semantic category acquisition module 830, and a capability index determination module 840.

[0124] Specifically, the keyword weight acquisition module 810 is used to acquire the weight of the keywords based on the keywords in the employee's script document;

[0125] Keyword scoring determination module 820 is used to determine the keyword scores in the employee script document based on the weight of the keywords;

[0126] The semantic category acquisition module 830 is used to classify the employee script document and obtain the semantic category score of the employee script document;

[0127] The competency indicator determination module 840 is used to determine the employee's competency indicators based on the keyword scores and semantic category scores in the employee's script document.

[0128] This script document processing device 800 can evaluate script documents and analyze employee capabilities based on the evaluation results. For example, it obtains the weight of keywords in an employee's script document, determines the keyword score based on the keyword weight, and categorizes the script document to obtain a semantic category score. Then, based on the keyword score and semantic category score, it determines the employee's capability indicators. This method allows for quick and accurate analysis of employee capability indicators, facilitating the development of subsequent business strategies based on those indicators.

[0129] According to an embodiment of the present invention, the script document processing device 800 can be used to implement... Figure 2 The implementation example describes a method for processing script documents.

[0130] In some embodiments of the present invention, the keyword weight acquisition module 810 is further configured to: extract the keywords from the employee script documents according to the keyword thesaurus; determine the weight of the keywords according to the mapping table between keywords and weights, wherein the weight of the keywords is the degree of difference of the keywords in multiple batches of historical script documents.

[0131] In some embodiments of the present invention, the keyword scoring determination module 820 is further configured to: calculate the sum of keyword weights based on the weight of each keyword among the plurality of keywords; and determine the keyword score in the employee script document based on the sum of keyword weights.

[0132] In some embodiments of the present invention, the semantic category acquisition module 830 is further configured to: input the employee script document into a prediction model, acquire semantic information in the employee script document through the prediction model, wherein the prediction model is a BERT model or a neural network model; and determine the semantic category score of the employee script document based on the semantic information of the employee script document.

[0133] In some embodiments of the present invention, the capability index determination module 840 is further configured to: concatenate the keyword scores and semantic category scores of the employee's script document to obtain a feature vector of the employee's script document; input the feature vector of the employee's script document into a logistic regression model to obtain a comprehensive score of the employee's script document through the logistic regression model; and match the capability index corresponding to the employee based on the comprehensive score of the employee's script document.

[0134] Figure 9 This is a schematic diagram of yet another example of the speech document processing apparatus of the present invention.

[0135] like Figure 9 As shown, the script document processing device 800 also includes a word set acquisition module 850, a keyword difference determination module 860, and a thesaurus and mapping table acquisition module 870.

[0136] Specifically, the word set acquisition module 850 is used to acquire a word set based on the multiple batches of historical speech documents.

[0137] The keyword difference determination module 860 is used to determine the difference of each word in the word set in different batches of historical speech documents based on the inverse document frequency of each word in the word set in different batches of historical speech documents and the word frequency of each word in the word set in different batches of historical speech documents.

[0138] The terminology and mapping table acquisition module 870 is used to acquire the terminology of the keywords and the mapping table of the keywords and their weights based on the degree of difference of each word in the terminology set in different batches of historical speech documents.

[0139] The script document processing device 800 can add words with greater differences to the keyword lexicon based on the degree of difference of each word in the word set in different batches of historical script documents. This allows for the accurate and rapid extraction of highly distinctive keywords from employee script documents based on the lexicon.

[0140] According to an embodiment of the present invention, the script document processing device 800 can be used to implement... Figure 4 The implementation example describes a method for processing script documents.

[0141] Since each module of the script document processing device 800 in the example embodiment of the present invention can be used to implement the above 2~ Figure 7 The steps of the described script document processing method are shown in the example embodiment. Therefore, for details not disclosed in the device embodiment of the present invention, please refer to the above-described embodiment of the script document processing method of the present invention.

[0142] It is understandable that the keyword weight acquisition module 810, keyword score determination module 820, semantic category acquisition module 830, capability indicator determination module 840, word set acquisition module 850, keyword difference determination module 860, and the thesaurus and mapping table acquisition module 870 can be combined into one module, or any one of these modules can be split into multiple modules. Alternatively, at least some of the functions of one or more of these modules can be combined with at least some of the functions of other modules and implemented in one module. According to embodiments of the present invention, at least one of the keyword weight acquisition module 810, keyword scoring determination module 820, semantic category acquisition module 830, capability index determination module 840, word set acquisition module 850, keyword difference determination module 860, and lexicon and mapping table acquisition module 870 can be at least partially implemented as hardware circuits, such as field-programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems-on-a-chip, systems-on-a-substrate, systems-on-package, application-specific integrated circuits (ASICs), or any other reasonable way of integrating or packaging circuits, or as hardware or firmware implementations, or as appropriate combinations of software, hardware, and firmware implementations. Alternatively, at least one of the keyword weight acquisition module 810, keyword scoring determination module 820, semantic category acquisition module 830, capability index determination module 840, word set acquisition module 850, keyword difference determination module 860, and lexicon and mapping table acquisition module 870 can be at least partially implemented as computer program modules, which, when run by a computer, can execute the functions of the corresponding modules.

[0143] The following describes embodiments of the computer device of the present invention, which can be considered as specific implementations of the methods and apparatus embodiments of the present invention described above. Details described in the computer device embodiments of the present invention should be considered as supplements to the methods or apparatus embodiments described above; details not disclosed in the computer device embodiments of the present invention can be implemented with reference to the methods or apparatus embodiments described above.

[0144] Figure 10This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. The computer device includes a processor and a memory. The memory is used to store a computer-executable program. When the computer program is executed by the processor, the processor performs the method described in any one of the embodiments, including but not limited to... Figure 2 The method.

[0145] like Figure 10 As shown, the computer device is represented in the form of a general-purpose computing device. There can be one or more processors working collaboratively. This invention also does not exclude distributed processing, meaning that processors can be distributed across different physical devices. The computer device of this invention is not limited to a single entity, but can also be the sum of multiple physical devices.

[0146] The memory stores a computer-executable program, typically machine-readable code. The computer-readable program can be executed by the processor to enable the computer device to perform the method of the present invention, or at least some steps of the method.

[0147] The memory includes volatile memory, such as random access memory (RAM) and / or cache memory, and may also be non-volatile memory, such as read-only memory (ROM).

[0148] Optionally, in this embodiment, the computer device further includes an I / O interface for exchanging data with external devices. The I / O interface can represent one or more of several bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0149] It should be understood that Figure 10 The computer device shown is merely an example of the present invention, and the computer device of the present invention may also include elements or components not shown in the above examples. For example, some computer devices also include display units such as screens, and some computer devices also include human-computer interaction elements such as buttons and keyboards. Any computer device capable of executing a computer-readable program in memory to implement the method of the present invention or at least some steps of the method can be considered as a computer device covered by the present invention.

[0150] Figure 11 This is a schematic diagram of a computer program product according to an embodiment of the present invention. Figure 11As shown, the computer program product stores a computer-executable program, which, when executed, implements the method described above. The computer program product may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer program product may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer program product may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0151] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0152] From the above description of the embodiments, those skilled in the art will readily understand that the present invention can be implemented by hardware capable of executing specific computer programs, such as the system of the present invention, and the electronic processing unit, server, client, mobile phone, control unit, processor, etc. included in the system. The present invention can also be implemented by computer software executing the methods of the present invention, for example, by control software executed by a microprocessor, electronic control unit, client, server, etc. However, it should be noted that the computer software executing the methods of the present invention is not limited to execution in one or a specific set of hardware entities; it can also be implemented in a distributed manner by unspecified hardware. For computer software, the software product can be stored in a computer-readable storage medium (such as a CD-ROM, USB flash drive, portable hard drive, etc.) or distributed across a network, as long as it enables computer devices to execute the methods according to the present invention.

[0153] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the present invention is not inherently related to any specific computer, virtual device, or computer equipment, and various general-purpose devices can also implement the present invention. The above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for processing script documents, characterized in that, include: Based on multiple batches of historical script documents, a vocabulary set was obtained; Based on the inverse document frequency and word frequency of each word in the word set in different batches of historical speech documents, determine the degree of difference of each word in the word set in different batches of historical speech documents; A keyword thesaurus was obtained from multiple batches of historical script documents. Each word in the employee's script documents was traversed to filter out words that were the same as the keywords in the thesaurus and extracted as keywords in the employee's script documents. Based on the keywords in the employee script documents, query the keyword-weight mapping table and obtain the weight corresponding to each keyword. The weight corresponding to each keyword is the degree of difference of the keyword in different historical script documents. If an employee's script contains multiple keywords, the weight of each keyword in the script is summed, and the keyword score in the script is determined by the sum of the calculated weights. The higher the keyword score, the better the selected keywords express the core content of the employee's script. The employee script documents are classified using a pre-trained prediction model to obtain multiple category label probabilities. The positive category label probabilities are used as the semantic category score of the employee script documents. The employee's competency indicators are determined based on the keyword scores and semantic category scores in the employee's script documents.

2. The script document processing method according to claim 1, characterized in that, The keyword database, derived from multiple batches of historical script documents, also includes: Based on the degree of difference of each word in the word set in different batches of historical speech documents, the keyword lexicon and the mapping table between keywords and weights are obtained.

3. The script document processing method according to claim 2, characterized in that, Also includes: A large degree of difference indicates that the meaning of the keywords differs significantly across different historical script documents; conversely, a small degree of difference indicates that the meaning of the keywords differs slightly across different historical script documents.

4. The script document processing method according to claim 3, characterized in that, Also includes: The differential keywords in different batches of historical speech documents are obtained by sorting the degree of difference of each word in the word set. A keyword thesaurus is built based on these differential keywords, and a mapping table between keywords and weights is built based on the differential keywords and their corresponding degrees of difference.

5. The script document processing method according to claim 1, characterized in that, The employee script documents are classified using a pre-trained prediction model to obtain multiple category label probabilities. The positive category label probabilities are used as the semantic category score of the employee script documents, including: Pre-trained prediction models are built and trained using BERT models or neural network models; Input the employee's script into the prediction model and use the prediction model to obtain the semantic information in the employee's script; Based on the semantic information of the employee script documents, determine the semantic category score of the employee script documents.

6. The script document processing method according to claim 1, characterized in that, Based on keyword scoring and semantic category scoring of employee communication scripts, the employee's competency indicators are determined as follows: The keyword scores and semantic category scores of the employee's script documents are concatenated to obtain the feature vector of the employee's script documents. The feature vectors of employee script documents are input into a logistic regression model, and the comprehensive score of the employee script documents is obtained through the logistic regression model. Based on the overall score of the employee's communication script, match the corresponding competency indicators with the employee.

7. The script document processing method according to claim 6, characterized in that, The overall score of employee communication scripts is obtained through a logistic regression model, including: The overall score of employee communication documents is calculated using the logistic regression algorithm in the logistic regression model. The logistic regression algorithm is as follows: Where x is the feature vector of the employee's script document, and w and b are the parameters in the logistic regression model.

8. A speech script document processing device, characterized in that, include: The keyword weight acquisition module is used to obtain a set of words based on multiple batches of historical script documents; Based on the inverse document frequency and word frequency of each word in the word set in different batches of historical script documents, the degree of difference of each word in the word set in different batches of historical script documents is determined. A keyword thesaurus is obtained from multiple batches of historical script documents. Each word in the employee script documents is traversed to filter out words that are the same as keywords in the thesaurus and extract them as keywords in the employee script documents. Based on the keywords in the employee script documents, the weight corresponding to each keyword is retrieved from the keyword-weight mapping table. The weight corresponding to each keyword is the degree of difference of the keyword in different historical script documents. The keyword scoring module is used to calculate the weight of each keyword in an employee's script document if the document contains multiple keywords. The calculated weight sum is then used to determine the keyword score in the employee's script document. The higher the keyword score, the better the selected keywords express the core content of the employee's script document. The semantic category acquisition module is used to classify employee script documents using a pre-trained prediction model, obtain multiple category label probabilities, and use the positive category label probability as the semantic category score of the employee script document. The competency indicator determination module is used to determine the competency indicators of employees based on keyword scores and semantic category scores in the employee's script documents.

9. A computer device comprising a processor and a memory, the memory being used to store a computer-executable program, characterized in that: When the computer program is executed by the processor, the processor performs the script document processing method as described in any one of claims 1-7.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the script document processing method according to any one of claims 1-7.

Images