User demand acquisition method and device, electronic equipment and storage medium

Abstract

The application discloses a user demand acquisition method, and relates to the fields of text recognition and user management. The method comprises the following steps: acquiring user generated content (UGC) text of a target data source; determining the text type of the UGC text by using a text classification model; wherein the text type comprises demand text, and the demand text describes a task to be completed by a user; if the UGC text is determined as the demand text by the text classification model, providing the demand text to a text generation model, so that the text generation model extracts task information corresponding to the demand text, and outputs a task object based on a specified sentence pattern and the task information. The technical scheme of the embodiment of the application not only realizes automatic classification of the UGC text, ensures screening and acquisition of the demand text in the UGC text, but also realizes extraction of task information and output of a structured task object, ensures the completeness and accuracy of the user demand acquisition result, and improves the acquisition efficiency of the user demand.

Description

Technical Field

[0001] This invention relates to the fields of text recognition and user management, and in particular to a method, apparatus, electronic device, and storage medium for obtaining user needs. Background Technology

[0002] As the types of products and services on various online platforms continue to increase, users' product and service needs are also constantly changing. How to effectively understand users' actual needs has become an important issue in the fields of system operation and maintenance and user management.

[0003] In existing technologies, the statistics and management of user needs are typically accomplished through manual surveys, sentiment analysis tools based on Neuro-Linguistic Programming (NLP), or knowledge graphs. Traditional manual surveys collect user needs through methods such as questionnaires, and then manually analyze and organize user feedback to form user need reports. Traditional sentiment analysis tools categorize user comments into positive, negative, and neutral sentiments to obtain user need information. Traditional knowledge graphs extract entities using manually defined entity types and rule models to obtain user needs.

[0004] However, traditional manual surveys require significant manpower and time, resulting in low efficiency in acquiring user needs; traditional sentiment analysis tools only assess surface-level polarity and fail to identify deeper user needs, often leading to substantial errors in statistical results; and traditional knowledge graphs can only perform statistical analysis on defined entity types, failing to capture comprehensive user needs. Summary of the Invention

[0005] This invention provides a method, apparatus, electronic device, storage medium, and computer program product for obtaining user needs, in order to solve the problems of being unable to obtain comprehensive user needs and low efficiency in obtaining user needs.

[0006] According to another aspect of the present invention, a method for obtaining user needs is provided, comprising: Obtain user-generated content (UGC) text from the target data source; The text type of the UGC text is determined by a text classification model; wherein, the text type includes requirement text, which describes a task to be completed by the user; If the UGC text is identified as a demand text by the text classification model, the demand text is provided to the text generation model so that the text generation model can extract the task information corresponding to the demand text and output the task object based on the specified sentence structure and the task information.

[0007] The task information includes at least one of contextual information, action information, and motivational information.

[0008] The method further includes: for multiple demand texts, performing clustering and deduplication processing on the multiple demand texts based on the semantic similarity between the task objects of each demand text to obtain clustered demand texts; providing the clustered demand texts to a sentiment evaluation model to obtain a sentiment score for the clustered demand texts through the sentiment evaluation model; wherein, the sentiment score includes a current push score, an ideal push score, a current resistance score, and an ideal resistance score; obtaining a driving force score for the clustered demand texts based on the sentiment score, and determining the demand priority of the clustered demand texts based on the driving force score.

[0009] The step of providing the clustering requirement text to the sentiment evaluation model to obtain the sentiment score of the clustering requirement text through the sentiment evaluation model includes: obtaining the central requirement text of the clustering requirement text and inputting the central requirement text to the sentiment evaluation model to obtain the sentiment score of the clustering requirement text through the sentiment evaluation model.

[0010] After determining the demand priority of the clustering demand text based on the driving force score, the method further includes: obtaining the external relationship label of each clustering demand text based on the semantic relationship between the multiple clustering demand texts; constructing a task graph based on the external relationship label and demand priority of each clustering demand text; wherein the clustering demand text serves as a task node in the task graph.

[0011] After constructing the task graph based on the external relationship labels and requirement priorities of each clustering requirement text, the method further includes: obtaining the internal relationship labels between each requirement text within each task node; and obtaining the subtask graph of each task node in the task graph based on the internal relationship labels between each requirement text.

[0012] After determining the text type of the UGC text using a text classification model, the method further includes: if the UGC text is not determined to be a demand text by the text classification model, obtaining the intent score of the UGC text through a predefined intent dictionary and obtaining the syntactic score of the UGC text through a predefined task sentence library; wherein, the intent dictionary includes multiple intent words; the task sentence library includes multiple task sentence patterns; and the text type of the UGC text is obtained based on the intent score and the syntactic score of the UGC text.

[0013] According to another aspect of the present invention, a user demand acquisition device is provided, comprising: The text acquisition module is used to acquire user-generated content (UGC) text from the target data source; The text type acquisition module is used to determine the text type of the UGC text through a text classification model; wherein, the text type includes requirement text, which describes a task to be completed by the user; The task object acquisition module is used to provide the requirement text to the text generation model if the UGC text is determined to be requirement text by the text classification model, so that the text generation model can extract the task information corresponding to the requirement text and output the task object based on the specified sentence structure and the task information.

[0014] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the user requirement acquisition method described in any embodiment of the present invention.

[0015] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the user requirement acquisition method described in any embodiment of the present invention.

[0016] According to another aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the user requirement acquisition method described in any embodiment of the present invention.

[0017] The technical solution of this invention involves acquiring user-generated content (UGC) text from a target data source; determining the text type of the UGC text using a text classification model; and providing the text generation model with the required text if the UGC text is identified as such by the text classification model. This allows the text generation model to extract the task information corresponding to the required text and output a task object based on a specified sentence structure and task information. This not only achieves automatic classification of UGC text, ensuring the selection and acquisition of required text within it, but also enables task information extraction and structured task object output, ensuring the comprehensiveness and accuracy of user requirement acquisition results and improving the efficiency of user requirement acquisition.

[0018] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a user requirement acquisition method provided according to Embodiment 1 of the present invention; Figure 2 This is a flowchart of another user requirement acquisition method provided according to Embodiment 2 of the present invention; Figure 3 This is a flowchart of another user requirement acquisition method provided in Embodiment 3 of the present invention; Figure 4 This is a flowchart of another user requirement acquisition method provided in Embodiment 4 of the present invention; Figure 5 This is a schematic diagram of a user demand acquisition device according to Embodiment 5 of the present invention; Figure 6 This is a schematic diagram of the structure of an electronic device that implements the user requirement acquisition method of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] Example 1 Figure 1This is a flowchart of a user requirement acquisition method provided in Embodiment 1 of the present invention. This embodiment is applicable to situations where task information is extracted from requirement text through a text generation model and a structured task object is output. This method can be executed by a user requirement acquisition device, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes: S101. Obtain user-generated content (UGC) text from the target data source.

[0024] User-generated content (UGC) refers to content information created and published by users on network platforms or business systems. UGC text can be published information presented in text form, or text information extracted from published information such as pictures, videos, and audio. The target data source, that is, the data source of UGC text, can come from one or more network platforms or business systems. Optionally, in the embodiments of the present invention, the type and source of UGC text are not specifically limited.

[0025] S102. Determine the text type of the UGC text through a text classification model; wherein, the text type includes requirement text, which describes a task to be completed by the user.

[0026] Text classification models can be machine learning models, such as classification models built and pre-trained based on techniques like Support Vector Machine (SVM), Random Forest, Convolutional Neural Network (CNN), and Recurrent Neural Network (RNN); or they can be large language models (LLM) based on deep learning, which intelligently classify text by understanding its context and semantic relationships.

[0027] The text classification model in this embodiment of the invention can be an existing general-purpose model on a network platform, accessed by calling a general interface on the network platform; alternatively, a text classification project can be created within the aforementioned general-purpose model, trained by uploading labeled samples, and the model can be deployed, thereby obtaining usage rights for the deployed model through a dedicated interface; alternatively, the aforementioned general-purpose model can be downloaded locally and fine-tuned locally using labeled samples, thus achieving local training of the text classification model; wherein, the labeled samples are labeled with classification categories including required text and non-required text.

[0028] Requirements text refers to text in which users clearly express their expectations, suggestions for improvement, or functional demands; for example, text in which users explicitly mention product functional defects or new requirements, text in which solutions are proposed based on usage scenarios, and text in which product advantages, disadvantages, or improvement plans are mentioned; non-requirements text refers to text that only covers emotional venting, irrelevant information, or invalid content; the user's job to be done (JBTD) described in the requirements text refers to the purpose or goal that the user wants to achieve through a product or service, which reflects the user's action intention from the user's perspective.

[0029] S103. If the UGC text is determined to be a demand text by the text classification model, the demand text is provided to the text generation model so that the text generation model can extract the task information corresponding to the demand text and output the task object based on the specified sentence structure and the task information.

[0030] The text generation model can be a large language model based on deep learning. It identifies the elements of the user's task to be completed in the requirement text through semantic parsing, such as task elements like objects, time, and actions. At the same time, by configuring a specified output sentence pattern for the text generation model, it guides the text generation model to output the extracted task information as a structured task object in the specified sentence pattern. This task object can express the task information in the requirement text in the form of a declarative sentence.

[0031] For example, if the requirement text is "Submit a summary report of Product A every Friday", the task information extracted by the text generation model includes action (submission), object (summary report of Product A), and time (every Friday), and the specified sentence structure is "periodic task: {action}; content: {object}; frequency: {time}", then the task object corresponding to the requirement text can be "periodic task: {submission}; content: {summary report of Product A}; frequency: {every Friday}".

[0032] Optionally, in this embodiment of the invention, the task information includes at least one of context information, action information, and motivation information. Context information represents the scene, time, or conditions of task execution, and can specifically be physical, social, or internal context information; action information represents the specific actions performed during task execution; motivation information represents the reason for performing the task or the desired goal, and can specifically include push forces (e.g., push forces generated by pain points) and multiple pull forces (i.e., pull forces generated by functional, emotional, or social factors).

[0033] After extracting contextual, action, and motivational information, the text generation model can output the task object in the specified sentence format "When [Context], I want to [Task], so I can [Motivation]". By extracting contextual, action, and motivational information, the model not only unifies the task information output format for texts with different needs, but also ensures a clear structure and logical closed loop in the extraction process, thereby improving the comprehensiveness and accuracy of the task object's information.

[0034] The technical solution of this invention involves acquiring user-generated content (UGC) text from a target data source; determining the text type of the UGC text using a text classification model; and providing the text generation model with the required text if the UGC text is identified as such by the text classification model. This allows the text generation model to extract the task information corresponding to the required text and output a task object based on a specified sentence structure and task information. This not only achieves automatic classification of UGC text, ensuring the selection and acquisition of required text within it, but also enables task information extraction and structured task object output, ensuring the comprehensiveness and accuracy of user requirement acquisition results and improving the efficiency of user requirement acquisition.

[0035] Example 2 Figure 2 This is a flowchart of a user requirement acquisition method provided in Embodiment 2 of the present invention. The relationship between this embodiment and the above embodiments is that, after UGC text is determined to be non-requirement text by a text classification model, the text is further classified using a predefined intent dictionary and task sentence library, such as... Figure 2 As shown, the method specifically includes: S201. Obtain user-generated content (UGC) text from the target data source.

[0036] S202. Determine the text type of the UGC text through a text classification model; wherein, the text type includes requirement text, which describes a task to be completed by the user.

[0037] S203. If the UGC text is determined to be a demand text by the text classification model, the demand text is provided to the text generation model so that the text generation model can extract the task information corresponding to the demand text and output the task object based on the specified sentence structure and the task information.

[0038] S204. If the UGC text is not identified as a demand text by the text classification model, the intent score of the UGC text is obtained through a predefined intent dictionary, and the syntactic score of the UGC text is obtained through a predefined task sentence library; wherein, the intent dictionary includes multiple intent words; the task sentence library includes multiple task sentence patterns.

[0039] Based on the word frequency statistics of historical requirement texts, an intent dictionary and task sentence library are pre-configured. The intent dictionary defines multiple strong intent words and weak intent words. Strong intent words are those that express the user's direct needs or action instructions and clearly reflect the intent, such as intending, planning, wanting to go, preparing, asking for recommendations, and needing. Weak intent words are those that express the user's initial interest or exploration needs through vague and broad words, without clearly pointing to specific actions or results, such as being conflicted, hesitant, and not knowing.

[0040] Each strong intent word and weak intent word corresponds to a different intent score, and the intent score of a strong intent word is greater than that of a weak intent word. For UGC texts that are not identified as demand texts by the text classification model, the number of strong intent words and weak intent words in the text is obtained, and the intent scores corresponding to each strong intent word and weak intent word are summed. The final sum is used as the intent score of the UGC text.

[0041] The task sentence library predefines several typical task sentence patterns and several negative sentence patterns. For example, "[first person] + [intentional verb] + [noun / verb]" is a typical task sentence pattern; while a sentence pattern containing words such as "I don't intend" or "it's unnecessary" is a negative sentence pattern. For UGC text that is not identified as demand text by the text classification model, it is matched with various sentence patterns in the task sentence library using regular expressions. If the sentence pattern of the current UGC text matches any task sentence pattern in the task sentence library, the syntactic score of the current UGC text is configured as a positive number; if the sentence pattern of the current UGC text does not match any sentence pattern in the task sentence library, the syntactic score of the current UGC text is configured as 0; if the sentence pattern of the current UGC text matches any negative sentence pattern in the task sentence library, the syntactic score of the current UGC text is configured as a negative number.

[0042] S205. Obtain the text type of the UGC text based on the intent score and the syntax score of the UGC text.

[0043] The evaluation score for each UGC text can be calculated based on its intent score and syntactic score, as well as the weights corresponding to the intent score and syntactic score. If the evaluation score of the current UGC text is greater than or equal to the preset evaluation threshold, it indicates that the current UGC text is a demand text; if the evaluation score of the current UGC text is less than the preset evaluation threshold, it indicates that the current UGC text is a non-demand text.

[0044] The technical solution of this invention, after the text classification model classifies UGC text into non-demand text, performs secondary classification of the UGC text from the user intent dimension and syntactic pattern dimension through a predefined intent dictionary and task sentence library. This avoids the text classification model from misclassifying some UGC text, especially UGC text with confidence close to the critical threshold, and greatly improves the accuracy of text classification results. It avoids both the omission of demand text and the misselection of non-demand text.

[0045] Example 3 Figure 3 This is a flowchart of a user requirement acquisition method provided in Embodiment 3 of the present invention. The relationship between this embodiment and the above embodiments is that the requirement text is clustered according to the task object, and the sentiment score of the clustered requirement text is obtained through a sentiment evaluation model, such as... Figure 3 As shown, the method specifically includes: S301. For multiple requirement texts, clustering and deduplication processing is performed on the multiple requirement texts according to the semantic similarity between the task objects of each requirement text to obtain clustered requirement texts.

[0046] To assess the semantic similarity between various requirement texts, the task objects of each requirement text can be used as comparison objects. The semantic similarity is obtained by calculating the Euclidean distance between each task object. Requirement texts with semantic similarity exceeding a preset similarity threshold are clustered, and requirement texts with identical text content are deduplicated to obtain the results, i.e., the clustered requirement texts. Alternatively, a large language model with embedded semantic similarity can be used to calculate the semantic similarity between various requirement texts, and clustering can be completed based on the semantic similarity calculation results.

[0047] S302. The clustering requirement text is provided to the sentiment evaluation model to obtain the sentiment score of the clustering requirement text through the sentiment evaluation model; wherein, the sentiment score includes the current driving force score, the ideal driving force score, the current resistance score, and the ideal resistance score.

[0048] The current state push reflects the degree of user dissatisfaction with the status quo, which is related to the intensity of negative emotions and the frequency of pain points. Specifically, it is the strength of the user's willingness to make new decisions driven by their dissatisfaction with the status quo. The ideal push reflects the attractiveness of the new solution to the user. Specifically, it is the strength of the user's willingness to make new decisions driven by their interest in the new solution.

[0049] Status quo resistance reflects users' inertia regarding the status quo, and is related to users' satisfaction with existing solutions, specifically the degree of satisfaction with existing solutions and the strength of their willingness to prevent users from making new decisions. Ideal resistance reflects users' anxiety about new solutions, and is related to users' concern about the consequences of changes, specifically the concern about new solutions and the strength of their willingness to prevent users from making new decisions.

[0050] As a generative large language model, the sentiment evaluation model can take the various requirement texts and / or task objects under the aggregated requirement text as input when inputting aggregated requirement text. It can also use prompt words to guide the sentiment evaluation model to extract the user's sentiment factors for the current task and score them from the above four aspects, namely, generating the current push force score, ideal push force score, current resistance score and ideal resistance score respectively.

[0051] For example, the prompt words for the configured sentiment evaluation model could be: "Please analyze the user's sentiment towards [Task] in the following text, and score (1-10 points) and explain the reasons based on the following four aspects: Current situation push: How dissatisfied the user is with the current situation; Ideal push: How much the user desires the new task; Current situation resistance: The resistance the user faces to changing the current situation; Ideal resistance: The user's concerns about the new task." S303. Based on the sentiment score of the clustering requirement text, obtain the driving force score of the clustering requirement text, and determine the requirement priority of the clustering requirement text based on the driving force score.

[0052] The sum of the current thrust score and the ideal thrust score is used as the thrust score, the sum of the current resistance score and the ideal resistance score is used as the resistance score, and the difference between the thrust score and the resistance score is used as the driving force score of the current demand text. Then, the demand priority of each cluster demand text is determined based on the driving force score. The driving force score and demand priority are positively correlated, that is, the larger the driving force score, the higher the demand priority. The demand priority also reflects the value of the current cluster demand text. The higher the demand priority, the greater the value.

[0053] Optionally, in this embodiment of the invention, providing the clustering requirement text to the sentiment evaluation model to obtain the sentiment score of the clustering requirement text through the sentiment evaluation model includes: obtaining the central requirement text of the clustering requirement text and inputting the central requirement text to the sentiment evaluation model to obtain the sentiment score of the clustering requirement text through the sentiment evaluation model.

[0054] Specifically, for clustered requirement texts, they are actually aggregated from multiple requirement texts with similar semantics. Under the current aggregated requirement text, the sum of the semantic similarity between each requirement text and other requirement texts is obtained, and the requirement text with the largest sum is determined as the aggregation center of the clustered requirement text, i.e., the central requirement text. Then, the central requirement text and the task object of the central requirement text are input into the sentiment evaluation model to obtain the sentiment score of the central requirement text, and the sentiment score of the central requirement text is used as the sentiment score of the current clustered requirement text. Compared with inputting each requirement text in the clustered requirement text into the sentiment evaluation model, inputting the most representative central requirement text in the clustered requirement text into the sentiment evaluation model greatly reduces the amount of text processing in the sentiment evaluation model and improves the efficiency of obtaining the sentiment score.

[0055] The technical solution of this invention addresses multiple requirement texts by clustering and deduplicating them based on the semantic similarity between the task objects of each text, thereby obtaining clustered requirement texts. These clustered requirement texts are then provided to a sentiment evaluation model to obtain a sentiment score. Based on the sentiment score, a driving force score is obtained for each clustered requirement text, and the priority of the requirement texts is determined according to this driving force score. Thus, the priority of each clustered requirement text reflects the value of that type of text, thereby achieving value assessment of different requirement texts and the selection of high-value requirement texts.

[0056] Example 4 Figure 4 This is a flowchart of a user requirement acquisition method provided in Embodiment 4 of the present invention. The relationship between this embodiment and the above embodiments is that a task graph is constructed based on the relationship tags between various clustered requirement texts, such as... Figure 4 As shown, the method specifically includes: S401. For multiple requirement texts, clustering and deduplication processing is performed on the multiple requirement texts according to the semantic similarity between the task objects of each requirement text to obtain clustered requirement texts.

[0057] S402. The clustering requirement text is provided to the sentiment evaluation model to obtain the sentiment score of the clustering requirement text through the sentiment evaluation model; wherein, the sentiment score includes the current push score, the ideal push score, the current resistance score, and the ideal resistance score.

[0058] S403. Based on the sentiment score of the clustering requirement text, obtain the driving force score of the clustering requirement text, and determine the requirement priority of the clustering requirement text based on the driving force score.

[0059] S404. Based on the semantic relationships between the multiple clustering requirement texts, obtain the external relationship labels of each clustering requirement text.

[0060] Since the clustering requirement text is composed of multiple semantically similar requirement texts, and the central requirement text reflects the central meaning of the clustering result, the central requirement text can be used as the basis for obtaining the relationship between the clustering requirement texts. By comparing the task objects of each central requirement text through a large language model, the relationship labels between each central requirement text can be identified, that is, the external relationship labels between each clustering requirement text. Among them, the relationships defined in the relationship labels can include hierarchical inclusion relationships, temporal order relationships, mutually exclusive conflict relationships, and unrelated relationships.

[0061] Hierarchical inclusion relationships indicate a hierarchical subordinate relationship between tasks. For example, clustering requirement text A is a subtask of clustering requirement text B. Temporal sequence relationships indicate a temporal dependency relationship between two tasks. For example, clustering requirement text A depends on the completion of clustering requirement text B, that is, clustering requirement text B must be completed before clustering requirement text A. Mutually exclusive relationships indicate that two tasks are mutually exclusive and cannot be executed simultaneously. No association relationship indicates that there is no defined relationship between two tasks.

[0062] S405. Construct a task graph based on the external relationship tags and requirement priorities of each clustering requirement text; wherein, the clustering requirement text serves as a task node in the task graph.

[0063] Each clustering requirement text is treated as a task node in the task graph. Based on the external relation labels of each clustering requirement text, the node relationships between each task node are established, and these node relationships exist in the task graph in the form of edges. At the same time, for each task node, the calculated requirement priority exists as node attribute information, thereby completing the construction of the task graph.

[0064] It is understood that in this embodiment of the invention, the task graph actually uses the demand text, i.e. the user's task to be completed, as nodes. The demand text is mined from massive UGC data. Therefore, it can capture changes in the user's task to be completed in a timely manner, i.e. changes in the user's related market demand. Moreover, the demand priority reflects the market value of the demand. Therefore, it can help enterprises to deeply understand consumers, discover emerging consumer groups, and provide a basis for market decision-making, product innovation, etc.

[0065] Optionally, in this embodiment of the invention, after constructing the task graph based on the external relationship tags and requirement priorities of each clustering requirement text, the method further includes: obtaining the internal relationship tags between each requirement text within each task node; and obtaining the subtask graph of each task node in the task graph based on the internal relationship tags between each requirement text.

[0066] Specifically, since each task node is actually an aggregated requirement text composed of multiple requirement texts, and there may be hierarchical inclusion relationships, temporal sequence relationships, and mutual exclusion and conflict relationships among the requirement texts under each task node, for each task node, the task objects between the internal requirement texts can be compared by a large language model to identify the relationship labels (i.e., internal relationship labels) between the internal requirement texts, and thus continue to build a subtask graph under each task node.

[0067] In particular, in the subtask graph, the demand priority of each demand text can be obtained based on the sentiment score of each demand text. This makes it easier to obtain higher-priority demand texts from the clustered demand texts with higher demand priority. Thus, by continuing to build subtask graphs within each task node of the task graph, the information integrity of the task graph is further improved.

[0068] Understandably, in another embodiment of the present invention, the relational tags of each clustering requirement text, including external relational tags and internal relational tags, can be obtained first, and then a task graph can be constructed based on the relational tags and requirement priorities of each clustering requirement text.

[0069] After constructing the task graph, the task graph can be output to display the task graph. Through the graph information such as nodes, relationships and requirement priorities, the graph viewer can be provided with market reference information such as the tasks to be completed by users and the value of the tasks, thereby providing a basis for market decision-making and product innovation.

[0070] The technical solution of this invention, after determining the demand priority of clustering demand texts based on the driving force score, further includes: obtaining external relationship tags for each clustering demand text based on the semantic relationships between multiple clustering demand texts; and constructing a task graph based on the external relationship tags and demand priorities of each clustering demand text. Thus, by using clustering demand texts, external relationship tags, and demand priorities as task nodes, edges, and node attribute information, respectively, the task graph is constructed. This task graph intuitively reflects the dependencies between various demand texts and the value of each demand text, achieving structured information generation based on all demand texts.

[0071] Example 5 Figure 5 This is a structural block diagram of a user demand acquisition device provided in Embodiment 5 of the present invention. The device specifically includes: Text acquisition module 501 is used to acquire user-generated content (UGC) text from the target data source; The text type acquisition module 502 is used to determine the text type of the UGC text through a text classification model; wherein, the text type includes requirement text, which describes a task to be completed by the user; The task object acquisition module 503 is used to provide the requirement text to the text generation model if the UGC text is determined to be requirement text by the text classification model, so that the text generation model can extract the task information corresponding to the requirement text and output the task object based on the specified sentence structure and the task information.

[0072] The technical solution of this invention involves acquiring user-generated content (UGC) text from a target data source; determining the text type of the UGC text using a text classification model; and providing the text generation model with the required text if the UGC text is identified as such by the text classification model. The model then extracts the task information corresponding to the required text and outputs a task object based on a specified sentence structure and task information. This not only achieves automatic classification of UGC text, ensuring the selection and acquisition of required text within it, but also enables task information extraction and structured task object output, ensuring the completeness and accuracy of user requirement acquisition results and improving the efficiency of user requirement acquisition.

[0073] Optionally, the task information includes at least one of contextual information, action information, and motivational information.

[0074] Optionally, the user demand acquisition device is further configured to perform clustering and deduplication processing on multiple demand texts based on the semantic similarity between the task objects of each demand text to obtain clustered demand texts; provide the clustered demand texts to a sentiment evaluation model to obtain a sentiment score for the clustered demand texts through the sentiment evaluation model; wherein the sentiment score includes a current push score, an ideal push score, a current resistance score, and an ideal resistance score; obtain a driving force score for the clustered demand texts based on the sentiment score, and determine the demand priority of the clustered demand texts based on the driving force score.

[0075] Optionally, the user demand acquisition device is further configured to acquire the central demand text of the clustered demand text and input the central demand text into the sentiment evaluation model to obtain the sentiment score of the clustered demand text through the sentiment evaluation model.

[0076] Optionally, the user requirement acquisition device is further configured to acquire external relationship tags of each clustering requirement text based on the semantic relationships between the multiple clustering requirement texts; and construct a task graph based on the external relationship tags and requirement priorities of each clustering requirement text; wherein the clustering requirement text serves as a task node in the task graph.

[0077] Optionally, the user requirement acquisition device is further configured to acquire internal relationship tags between each requirement text within each task node; and acquire the subtask graph of each task node in the task graph based on the internal relationship tags between each requirement text.

[0078] Optionally, the user demand acquisition device is further configured to, if the UGC text is not determined as demand text by the text classification model, obtain the intent score of the UGC text through a predefined intent dictionary and obtain the syntactic score of the UGC text through a predefined task sentence library; wherein, the intent dictionary includes multiple intent words; the task sentence library includes multiple task sentence patterns; and the text type of the UGC text is obtained based on the intent score and the syntactic score of the UGC text.

[0079] The above-described apparatus can execute the user requirement acquisition method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the user requirement acquisition method provided in any embodiment of the present invention.

[0080] Example 6 Figure 6A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, electronic devices, blade electronic devices, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0081] like Figure 6 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0082] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0083] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as user requirement retrieval methods.

[0084] In some embodiments, the user requirement acquisition method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on a heterogeneous hardware accelerator via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by a processor, one or more steps of the user requirement acquisition method described above may be performed. Alternatively, in other embodiments, the processor may be configured to perform the user requirement acquisition method by any other suitable means (e.g., by means of firmware).

[0085] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0086] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0087] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0088] To provide interaction with a user terminal, the systems and techniques described herein can be implemented on a heterogeneous hardware accelerator, which includes: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user terminal; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user terminal provides input to the heterogeneous hardware accelerator. Other types of devices can also be used to provide interaction with the user terminal; for example, the feedback provided to the user terminal can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or haptic feedback); and input from the user terminal can be received in any form (including sound input, voice input, or haptic input).

[0089] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., client computers with graphical user interfaces or web browsers through which client computers can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0090] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0091] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0092] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for obtaining user needs, characterized in that, include: Obtain user-generated content (UGC) text from the target data source; The text type of the UGC text is determined by a text classification model; wherein, the text type includes requirement text, which describes a task to be completed by the user; If the UGC text is identified as a demand text by the text classification model, the demand text is provided to the text generation model so that the text generation model can extract the task information corresponding to the demand text and output the task object based on the specified sentence structure and the task information.

2. The user demand acquisition method according to claim 1, characterized in that, The task information includes at least one of contextual information, action information, and motivational information.

3. The user demand acquisition method according to claim 1, characterized in that, The method further includes: For multiple requirement texts, clustering and deduplication processing is performed on the multiple requirement texts based on the semantic similarity between the task objects of each requirement text to obtain clustered requirement texts; The clustering requirement text is provided to the sentiment evaluation model to obtain the sentiment score of the clustering requirement text; wherein, the sentiment score includes the current push score, the ideal push score, the current resistance score, and the ideal resistance score. Based on the sentiment score of the clustering requirement text, the driving force score of the clustering requirement text is obtained, and the requirement priority of the clustering requirement text is determined based on the driving force score.

4. The user demand acquisition method according to claim 3, characterized in that, The step of providing the clustering requirement text to the sentiment rating model, so as to obtain the sentiment score of the clustering requirement text through the sentiment rating model, includes: Obtain the central requirement text of the clustering requirement text and input the central requirement text into the sentiment evaluation model to obtain the sentiment score of the clustering requirement text through the sentiment evaluation model.

5. The user demand acquisition method according to claim 3, characterized in that, After determining the demand priority of the clustering demand text based on the driving force score, the method further includes: Based on the semantic relationships between multiple clustering requirement texts, obtain the external relationship labels of each clustering requirement text; A task graph is constructed based on the external relationship tags and requirement priorities of each clustering requirement text; wherein, the clustering requirement text serves as a task node in the task graph.

6. The user demand acquisition method according to claim 5, characterized in that, After constructing the task graph based on the external relation labels and requirement priorities of each clustering requirement text, the process also includes: Obtain the internal relationship tags between the various requirement texts within each task node; Based on the internal relationship tags between the various requirement texts, obtain the subtask graph of each task node in the task graph.

7. The user demand acquisition method according to claim 1, characterized in that, After determining the text type of the UGC text using a text classification model, the process also includes: If the UGC text is not identified as a demand text by the text classification model, the intent score of the UGC text is obtained through a predefined intent dictionary, and the syntactic score of the UGC text is obtained through a predefined task sentence library; wherein, the intent dictionary includes multiple intent words; the task sentence library includes multiple task sentence patterns; The text type of the UGC text is obtained based on the intent score and the syntax score of the UGC text.

8. A user demand acquisition device, characterized in that, include: The text acquisition module is used to acquire user-generated content (UGC) text from the target data source; The text type acquisition module is used to determine the text type of the UGC text through a text classification model; wherein, the text type includes requirement text, which describes a task to be completed by the user; The task object acquisition module is used to provide the requirement text to the text generation model if the UGC text is determined to be requirement text by the text classification model, so that the text generation model can extract the task information corresponding to the requirement text and output the task object based on the specified sentence structure and the task information.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the user requirement acquisition method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the user requirement acquisition method according to any one of claims 1-7.

11. A computer program product comprising a computer program that, when executed by a processor, implements the user requirement acquisition method according to any one of claims 1-7.

Images