Text mining method, device and equipment and storage medium

Abstract

The present application relates to artificial intelligence, and provides a text mining method, device, equipment and storage medium. When it is monitored that the mining template is successfully triggered in the operation interface, the target vocabulary is generated according to the positive input vocabulary, the mining relationship is identified according to the input position of the target vocabulary in the operation interface, the search sentence is generated according to the word distance of the template vocabulary, the negative input vocabulary, the mining relationship and the target vocabulary, the search thread is called to execute the search sentence according to the text quantity of the template type in the preset micro-service architecture, the search text is efficiently and accurately obtained, and then the mining detail interface is generated. In addition, the present application also relates to blockchain technology, and the mining detail interface can be stored in the blockchain.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a text mining method, apparatus, device, and storage medium. Background Technology

[0002] In today's era of big data, industries such as finance have accumulated massive amounts of data, including both structured and unstructured text. As businesses interact with customers in increasingly diverse ways, much valuable information is hidden within this vast amount of interactive data. Therefore, the demand for text mining is growing daily among businesses.

[0003] In the field of artificial intelligence, there are relevant text mining solutions. However, due to the lack of structured tags (such as customer tags, agent profiles, business information, etc.) to help limit the mining scope, these solutions cannot efficiently and accurately extract valuable text information from massive amounts of data. Summary of the Invention

[0004] In view of the above, it is necessary to provide a text mining method, apparatus, device and storage medium that can solve the technical problem of being unable to efficiently and accurately mine valuable text information from massive amounts of information.

[0005] On the one hand, the present invention proposes a text mining method, the text mining method comprising:

[0006] When a text mining request is received, a mining template is displayed in the operation interface according to the text mining request;

[0007] When it is detected that the mining template is successfully triggered in the operation interface, target words are generated based on the positive input words in the operation interface. The target words include the positive input words and the keywords in the associated words corresponding to the positive input words.

[0008] The mining relationship is identified based on the input position of the target words in the operation interface;

[0009] A search statement is generated based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words;

[0010] Based on the number of texts in the preset microservice architecture corresponding to the template type of the mining template, the search thread in the preset search engine is invoked to execute the search statement and obtain the search text;

[0011] A mining details interface is generated based on the number of texts and the search text.

[0012] According to a preferred embodiment of the present invention, before generating target words based on the positive input words in the operation interface, the text mining method further includes:

[0013] Identify the initial display time of the excavation template on the operation interface;

[0014] Based on the initial display time, obtain the operation log of the operation interface;

[0015] Identify the button positions of the associated buttons related to the mining template from the operation interface, and identify the box positions of the text boxes in the operation interface;

[0016] Check whether the operation log includes the button position and the box position;

[0017] If the operation log includes the button position and / or the box position, then it is determined that the mining template was successfully triggered in the operation interface.

[0018] According to a preferred embodiment of the present invention, the frame position includes a positive position and a negative position, and the step of generating target words based on the positive input words in the operation interface includes:

[0019] Obtain the positive input vocabulary from the positive position;

[0020] Based on the pre-trained synonym prediction model, generate the associated words corresponding to the positive input words;

[0021] Check whether the associated words are successfully triggered;

[0022] The associated words that successfully trigger the association will be identified as the keywords.

[0023] The target vocabulary is generated based on the positive input vocabulary and the keywords. The box position includes positive and negative positions. Generating the target vocabulary based on the positive input vocabulary in the operation interface includes:

[0024] Obtain the positive input vocabulary from the positive position;

[0025] Based on the pre-trained synonym prediction model, generate the associated words corresponding to the positive input words;

[0026] Check whether the associated words are successfully triggered;

[0027] The associated words that successfully trigger the association will be identified as the keywords.

[0028] The target vocabulary is generated based on the positive input vocabulary and the keywords.

[0029] According to a preferred embodiment of the present invention, the synonym prediction model includes a semantic analysis network and a prediction output network. The semantic analysis network includes a forward feature extraction network and a backward feature extraction network. The generation of associated words corresponding to the positive input words based on the pre-trained synonym prediction model includes:

[0030] The positive input vocabulary is encoded to obtain the input vector;

[0031] Based on the positive feature extraction network, features are extracted from the input vector to obtain the first feature;

[0032] Based on the inverse feature extraction network, features are extracted from the input vector to obtain the second feature;

[0033] Generate semantic vectors for the positive input words based on the first feature and the second feature;

[0034] Based on the semantic vector, the corresponding words are obtained from the prediction output network as the associated words.

[0035] According to a preferred embodiment of the present invention, generating a search statement based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words includes:

[0036] Identify the word type of the template words in the mining template, and identify the word relationships of the template words in the mining template;

[0037] Generate negative search terms from the template terms of the negative term type and the negative input terms;

[0038] Generate positive search terms from the template terms of the positive term type and the target terms;

[0039] The word relationships, negative search terms, positive search terms, mining relationships, and word distances are written into a preset query statement to obtain the search statement.

[0040] According to a preferred embodiment of the present invention, the step of generating a mining details interface based on the number of texts and the search text includes:

[0041] Count the number of searches for the search text;

[0042] A search ratio is generated based on the ratio of the number of searches to the number of texts.

[0043] Count the frequency of each positive search term in the search text;

[0044] The matching degree between the search text and the search statement is calculated based on the frequency.

[0045] Based on the matching degree, the target text is selected from the multiple search texts;

[0046] The number of texts, the number of searches, the search ratio, the frequency, and the target text are written into a preset interface to obtain the mining details interface.

[0047] According to a preferred embodiment of the present invention, the step of calling a search thread in a preset search engine to execute the search statement based on the number of texts of the template type corresponding to the mining template in a preset microservice architecture, and obtaining the search text, includes:

[0048] Based on the template type, the public opinion texts stored in the preset microservice architecture are quantified to obtain the number of texts;

[0049] Obtain the processing rate of idle threads from the preset search engine;

[0050] Extract the mining time requirement from the text mining request;

[0051] The processing rate with the smallest value is determined as the target rate;

[0052] The number of threads is calculated based on the number of texts, the required mining time, and the target rate.

[0053] The search thread is selected from the plurality of idle threads based on the number of threads and the processing rate;

[0054] The search thread is invoked to compare the search statement with the public opinion text, and the public opinion text that successfully matches the search statement is identified as the search text.

[0055] On the other hand, the present invention also proposes a text mining device, the text mining device comprising:

[0056] The display unit is used to display a mining template in the operation interface when a text mining request is received;

[0057] The generation unit is used to generate target words based on the positive input words in the operation interface when the mining template is successfully triggered in the operation interface. The target words include the positive input words and the keywords in the associated words corresponding to the positive input words.

[0058] The identification unit is used to identify the mining relationship based on the input position of the target word in the operation interface;

[0059] The generation unit is further configured to generate search statements based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words;

[0060] An execution unit is used to call a search thread in a preset search engine to execute the search statement based on the number of texts of the template type corresponding to the mining template in a preset microservice architecture, so as to obtain the search text;

[0061] The generation unit is also used to generate a mining details interface based on the number of texts and the search text.

[0062] On the other hand, the present invention also proposes an electronic device, the electronic device comprising:

[0063] Memory, which stores computer-readable instructions; and

[0064] The processor executes computer-readable instructions stored in the memory to implement the text mining method.

[0065] On the other hand, the present invention also proposes a computer-readable storage medium storing computer-readable instructions, which are executed by a processor in an electronic device to implement the text mining method.

[0066] As can be seen from the above technical solutions, this application, by identifying the mining template corresponding to the text mining request, can quickly generate the search statement and improve the efficiency of search text generation because the user does not need to repeatedly input the template content in the operation interface. When the mining template is detected to be triggered, it indicates that the mining template needs further improvement. Furthermore, by using the positive input words, it can generate associated words to assist the user's input, thereby improving the efficiency of target word generation. Simultaneously, it can assist in locking the mining scope when the user lacks structured tags. Combining the template words, the negative input words, the mining relationships, and the word distances, it can accurately generate search statements, thereby improving the accuracy of search text generation. By using the number of texts, it can reasonably call the search thread to process the search statement, thus achieving rapid processing of the search text without affecting other request processing. In addition, this application also uses the number of texts and the search text to generate a mining details interface, which can help the user quickly understand the response status of the text mining request. Attached Figure Description

[0067] Figure 1 This is a flowchart of a preferred embodiment of the text mining method of the present invention.

[0068] Figure 2 This is a model structure diagram of the synonym prediction model in the text mining method of this invention.

[0069] Figure 3 This is a functional block diagram of a preferred embodiment of the text mining device of the present invention.

[0070] Figure 4 This is a schematic diagram of the structure of an electronic device that implements the text mining method of the present invention. Detailed Implementation

[0071] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0072] like Figure 1 The diagram shown is a flowchart of a preferred embodiment of the text mining method of the present invention. The order of the steps in this flowchart can be changed, and some steps can be omitted, depending on different requirements.

[0073] The text mining method described above can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0074] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0075] The text mining method is applied to one or more electronic devices, which are devices capable of automatically performing numerical calculations and / or information processing according to pre-set or stored computer-readable instructions. Their hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0076] The electronic device can be any electronic product that can interact with the user, such as a personal computer, tablet computer, smartphone, personal digital assistant (PDA), game console, interactive network television (IPTV), smart wearable device, etc.

[0077] The electronic devices may include network devices and / or user devices. The network devices include, but are not limited to, single network electronic devices, groups of multiple network electronic devices, or cloud computing-based systems consisting of a large number of hosts or network electronic devices.

[0078] The network in which the electronic device is located includes, but is not limited to: the Internet, wide area network, metropolitan area network, local area network, virtual private network (VPN), etc.

[0079] 101. When a text mining request is received, a mining template is displayed in the operation interface according to the text mining request.

[0080] In at least one embodiment of the present invention, the text mining request is a request generated in response to a user triggering a mining request button on the operation interface.

[0081] The user interface refers to the interface in the electronic device that can be operated by the user.

[0082] The mining template refers to a template that matches the scenario type corresponding to the requested scenario.

[0083] In at least one embodiment of the present invention, the electronic device displays a mining template in the user interface according to the text mining request, including:

[0084] Parse the request message of the text mining request to obtain the data information carried by the request message;

[0085] Extract the request scenario from the data information;

[0086] Identify the scenario type corresponding to the requested scenario as the template type;

[0087] Obtain a template corresponding to the template type from the preset template library as the mining template;

[0088] Based on the generation log of the text mining request, the trigger interface of the file mining request is located as the operation interface;

[0089] The mining template is displayed based on the user interface.

[0090] The data information includes the request scenario and the required mining time. The request scenario includes, but is not limited to, complaints and investigations, and special research projects.

[0091] The preset template library stores multiple template types and their mapping relationships.

[0092] The generated log refers to the log file generated when the text mining request is triggered.

[0093] The trigger interface refers to the interface where the "mining demand" button is located.

[0094] The request message allows for the rapid acquisition of the request scenario, which in turn enables the accurate acquisition of the mining template. Furthermore, the generated logs allow for the accurate location of the operation interface.

[0095] 102. When it is detected that the mining template is successfully triggered in the operation interface, target words are generated based on the positive input words in the operation interface. The target words include the positive input words and the keywords in the associated words corresponding to the positive input words.

[0096] In at least one embodiment of the present invention, the positive input vocabulary refers to the vocabulary entered by the user in a positive position on the operation interface. The positive input vocabulary refers to the vocabulary that must be included in the search text.

[0097] The associated vocabulary refers to synonyms of the forward input vocabulary.

[0098] The keywords refer to the associated words that are successfully triggered by the user.

[0099] In at least one embodiment of the present invention, before generating target words based on positive input words in the user interface, the text mining method further includes:

[0100] Identify the initial display time of the excavation template on the operation interface;

[0101] Based on the initial display time, obtain the operation log of the operation interface;

[0102] Identify the button positions of the associated buttons related to the mining template from the operation interface, and identify the box positions of the text boxes in the operation interface;

[0103] Check whether the operation log includes the button position and the box position;

[0104] If the operation log includes the button position and / or the box position, then it is determined that the mining template was successfully triggered in the operation interface.

[0105] The starting display time refers to the initial time when the excavation template is displayed on the operation interface.

[0106] The operation log refers to days when the log generation time is greater than the starting display time, and the operation log is the log corresponding to the operation interface.

[0107] The associated button can be a one-click application button, etc.

[0108] The operation log can be accurately obtained through the start display time and the operation interface. Furthermore, by identifying whether the operation log includes the button position and the box position, it is possible to accurately identify whether the mining template has been successfully triggered in the operation interface.

[0109] In other embodiments, if the operation log does not include the button position and the box position, it is determined that the mining template was not successfully triggered in the operation interface.

[0110] In at least one embodiment of the present invention, the frame position includes a positive position and a negative position, and the electronic device generates target words based on the positive input words in the operation interface, including:

[0111] Obtain the positive input vocabulary from the positive position;

[0112] Based on the pre-trained synonym prediction model, generate the associated words corresponding to the positive input words;

[0113] Check whether the associated words are successfully triggered;

[0114] The associated words that successfully trigger the association will be identified as the keywords.

[0115] The target vocabulary is generated based on the positive input vocabulary and the keywords.

[0116] The positive position refers to the position of words that must be included in the search text, and the negative position refers to the position of words that must not be included in the search text.

[0117] like Figure 2 The diagram shown is a model structure diagram of the synonym prediction model in the text mining method of this invention. Figure 2 In the above, the synonym prediction model includes a semantic analysis network and a prediction output network, and the semantic analysis network includes a forward feature extraction network and a reverse feature extraction network.

[0118] The synonym prediction model can generate associated words, thereby assisting users in text mining, improving the convenience of mining, and avoiding the inability to fully obtain search text due to the lack of structured tags.

[0119] Specifically, the electronic device generates associated words corresponding to the positive input words based on a pre-trained synonym prediction model, including:

[0120] The positive input vocabulary is encoded to obtain the input vector;

[0121] Based on the positive feature extraction network, features are extracted from the input vector to obtain the first feature;

[0122] Based on the inverse feature extraction network, features are extracted from the input vector to obtain the second feature;

[0123] Generate semantic vectors for the positive input words based on the first feature and the second feature;

[0124] Based on the semantic vector, the corresponding words are obtained from the prediction output network as the associated words.

[0125] The forward feature extraction network and the reverse feature extraction network each include convolutional layers.

[0126] The forward feature extraction network refers to a network that analyzes the vector elements of the input vector in a forward-to-back order. The reverse feature extraction network refers to a network that analyzes the vector elements of the input vector in a backward-to-forward order.

[0127] The semantic vector refers to the average value of each feature element in the first feature and the corresponding feature element in the second feature.

[0128] The prediction output network includes fully connected layers and a vector-vocabulary mapping table.

[0129] By combining the forward feature extraction network and the reverse feature extraction network to analyze the input vector, the semantic vector can be accurately generated. Then, the semantic vector can be accurately processed by the prediction output network, thereby improving the accuracy of generating the associated words.

[0130] 103. Identify the mining relationship based on the input position of the target word in the operation interface.

[0131] In at least one embodiment of the present invention, the input position includes a series position and a parallel position.

[0132] The mining relationships include serial relationships (i.e., and) and parallel relationships (i.e., or).

[0133] In at least one embodiment of the present invention, the electronic device identifies the mining relationship based on the input position of the target word in the operation interface, including:

[0134] The mining relationship corresponding to the target words at the aforementioned concatenation position is determined as a concatenation relationship;

[0135] The mining relationships corresponding to the target words in the parallel positions are determined as parallel relationships.

[0136] The mining relationship can be accurately identified through the input location.

[0137] 104. Generate a search statement based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words.

[0138] In at least one embodiment of the present invention, the template vocabulary refers to the vocabulary information stored in the mining template, and the template vocabulary includes template vocabulary with negative vocabulary type and template vocabulary with positive vocabulary type.

[0139] The negative input words refer to the words entered by the user in the negative position of the operation interface, and the negative input words refer to words that must not be included in the search text.

[0140] The word distance refers to the number of words between any two target words. For example, if the word distance is 2, the search text could be: "opened a company".

[0141] In at least one embodiment of the present invention, the electronic device generates a search statement based on template words in the mining template, negative input words in the operation interface, the mining relationship, and the word distance between the target words, including:

[0142] Identify the word type of the template words in the mining template, and identify the word relationships of the template words in the mining template;

[0143] Generate negative search terms from the template terms of the negative term type and the negative input terms;

[0144] Generate positive search terms from the template terms of the positive term type and the target terms;

[0145] The word relationships, negative search terms, positive search terms, mining relationships, and word distances are written into a preset query statement to obtain the search statement.

[0146] The lexical relationships include the serial relationships and the parallel relationships.

[0147] The preset query statement can be a structured query statement.

[0148] By identifying the word type, words of the negative type can be filtered from the template words, and negative search words can be generated comprehensively by combining the negative input words. By filtering words of the positive type from the template words, positive search words can be generated comprehensively by combining the positive input words. At the same time, the search statement can be generated quickly through the preset query statement, thereby improving the efficiency of text mining.

[0149] 105. Based on the number of texts in the preset microservice architecture corresponding to the template type of the mining template, the search thread in the preset search engine is invoked to execute the search statement to obtain the search text.

[0150] In at least one embodiment of the present invention, the template type includes, but is not limited to, non-renewal of insurance, etc.

[0151] The preset microservice architecture can be an architecture composed of multiple database servers.

[0152] The preset search engine stores multiple search threads.

[0153] The search query refers to the public opinion text that successfully matches the search query in the preset microservice architecture.

[0154] In at least one embodiment of the present invention, the electronic device, based on the number of texts of the template type corresponding to the mining template in a preset microservice architecture, calls a search thread in a preset search engine to execute the search statement, and obtains search text including:

[0155] Based on the template type, the public opinion texts stored in the preset microservice architecture are quantified to obtain the number of texts;

[0156] Obtain the processing rate of idle threads from the preset search engine;

[0157] Extract the mining time requirement from the text mining request;

[0158] The processing rate with the smallest value is determined as the target rate;

[0159] The number of threads is calculated based on the number of texts, the required mining time, and the target rate.

[0160] The search thread is selected from the plurality of idle threads based on the number of threads and the processing rate;

[0161] The search thread is invoked to compare the search statement with the public opinion text, and the public opinion text that successfully matches the search statement is identified as the search text.

[0162] The idle threads refer to all execution threads in the preset search engine that are in an idle state.

[0163] The mining time requirement refers to the maximum time period that can be allowed in response to the text mining request.

[0164] The formula for generating the number of threads is: Where n1 represents the number of threads, n2 represents the number of texts, v represents the target rate, and t represents the mining time requirement.

[0165] The search thread refers to the thread selected from the plurality of idle threads in descending order of processing speed, and the number of search threads is equal to the number of threads.

[0166] In this embodiment, since the text mining request is a further limited search on the mining template, the number of texts can be reasonably quantified by the template type. Then, by selecting the processing rate with the smallest value as the target rate, the number of threads can be determined as large as possible. Thus, by combining the number of threads with the processing rate, the search threads can be filtered out to ensure that the text mining request can be responded to within the mining time requirement.

[0167] In at least one embodiment of the present invention, when the public opinion text does not contain the negative search terms, contains the positive search terms, and satisfies the word distance, the electronic device determines the public opinion text as the search text.

[0168] 106. Generate a mining details interface based on the number of texts and the search text.

[0169] It should be emphasized that, to further ensure the privacy and security of the mining details interface, the mining details interface can also be stored in a blockchain node.

[0170] In at least one embodiment of the present invention, the mining details interface includes the number of texts, the number of searches for the search text, etc.

[0171] In at least one embodiment of the present invention, the electronic device generates a mining details interface based on the number of texts and the search text, including:

[0172] Count the number of searches for the search text;

[0173] A search ratio is generated based on the ratio of the number of searches to the number of texts.

[0174] Count the frequency of each positive search term in the search text;

[0175] The matching degree between the search text and the search statement is calculated based on the frequency.

[0176] Based on the matching degree, the target text is selected from the multiple search texts;

[0177] The number of texts, the number of searches, the search ratio, the frequency, and the target text are written into a preset interface to obtain the mining details interface.

[0178] The higher the frequency, the higher the matching degree.

[0179] The target text refers to the N search texts with the highest matching degree, and N can be set according to the interface size of the preset interface.

[0180] The preset interface can be a blank interface, or it can be an interface that includes the number of texts, the number of searches, the search ratio, the frequency, and the tag information of the target text.

[0181] By statistically analyzing the frequency of each positive search term, the matching degree can be accurately quantified. The matching degree can then be used to reasonably filter the target text, which is then displayed in the mining details interface, improving the intuitiveness of the target text.

[0182] As can be seen from the above technical solutions, this application, by identifying the mining template corresponding to the text mining request, can quickly generate the search statement and improve the efficiency of search text generation because the user does not need to repeatedly input the template content in the operation interface. When the mining template is detected to be triggered, it indicates that the mining template needs further improvement. Furthermore, by using the positive input words, it can generate associated words to assist the user's input, thereby improving the efficiency of target word generation. Simultaneously, it can assist in locking the mining scope when the user lacks structured tags. Combining the template words, the negative input words, the mining relationships, and the word distances, it can accurately generate search statements, thereby improving the accuracy of search text generation. By using the number of texts, it can reasonably call the search thread to process the search statement, thus achieving rapid processing of the search text without affecting other request processing. In addition, this application also uses the number of texts and the search text to generate a mining details interface, which can help the user quickly understand the response status of the text mining request.

[0183] like Figure 3 The diagram shown is a functional block diagram of a preferred embodiment of the text mining device of the present invention. The text mining device 11 includes a display unit 110, a generation unit 111, a recognition unit 112, an execution unit 113, a determination unit 114, an acquisition unit 115, and a detection unit 116. The module / unit referred to in this invention refers to a series of computer-readable instruction segments that can be acquired by the processor 13 and perform a fixed function, and are stored in the memory 12. In this embodiment, the functions of each module / unit will be described in detail in subsequent embodiments.

[0184] The display unit 110 is used to display a mining template in the operation interface according to the text mining request when a text mining request is received.

[0185] The generation unit 111 is used to generate target words based on the positive input words in the operation interface when it is detected that the mining template is successfully triggered in the operation interface. The target words include the positive input words and the keywords in the associated words corresponding to the positive input words.

[0186] The identification unit 112 is used to identify the mining relationship based on the input position of the target word in the operation interface;

[0187] The generation unit 111 is further configured to generate search statements based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words;

[0188] Execution unit 113 is used to call the search thread in the preset search engine to execute the search statement based on the number of texts of the template type corresponding to the mining template in the preset microservice architecture, so as to obtain the search text;

[0189] The generation unit 111 is also used to generate a mining details interface based on the number of texts and the search text.

[0190] In at least one embodiment of the present invention, before generating target words based on the positive input words in the operation interface, the recognition unit 112 is further configured to recognize the starting display time of the mining template on the operation interface;

[0191] The acquisition unit 115 is used to acquire the operation log of the operation interface based on the start display time;

[0192] The identification unit 112 is also used to identify the button position of the associated button associated with the mining template from the operation interface, and to identify the box position of the text box in the operation interface.

[0193] Detection unit 116 is used to detect whether the operation log includes the button position and the box position;

[0194] The determining unit 114 is used to determine that the mining template has been successfully triggered in the operation interface if the operation log includes the button position and / or the box position.

[0195] The operation log can be accurately obtained through the start display time and the operation interface. Furthermore, by identifying whether the operation log includes the button position and the box position, it is possible to accurately identify whether the mining template has been successfully triggered in the operation interface.

[0196] In other embodiments, the determining unit 114 is further configured to determine that the mining template has not been successfully triggered in the operation interface if the operation log does not include the button position and the box position.

[0197] In at least one embodiment of the present invention, the box position includes a positive position and a negative position, and the generation unit 111 is further configured to obtain the positive input words from the positive position;

[0198] Based on the pre-trained synonym prediction model, generate the associated words corresponding to the positive input words;

[0199] Check whether the associated words are successfully triggered;

[0200] The associated words that successfully trigger the association will be identified as the keywords.

[0201] The target vocabulary is generated based on the positive input vocabulary and the keywords.

[0202] The synonym prediction model can generate associated words, thereby assisting users in text mining, improving the convenience of mining, and avoiding the inability to fully obtain search text due to the lack of structured tags.

[0203] In at least one embodiment of the present invention, the synonym prediction model includes a semantic analysis network and a prediction output network. The semantic analysis network includes a forward feature extraction network and a reverse feature extraction network. The generation unit 111 is further used to encode the forward input words to obtain an input vector.

[0204] Based on the positive feature extraction network, features are extracted from the input vector to obtain the first feature;

[0205] Based on the inverse feature extraction network, features are extracted from the input vector to obtain the second feature;

[0206] Generate semantic vectors for the positive input words based on the first feature and the second feature;

[0207] Based on the semantic vector, the corresponding words are obtained from the prediction output network as the associated words.

[0208] By combining the forward feature extraction network and the reverse feature extraction network to analyze the input vector, the semantic vector can be accurately generated. Then, the semantic vector can be accurately processed by the prediction output network, thereby improving the accuracy of generating the associated words.

[0209] In at least one embodiment of the present invention, the generation unit 111 is further configured to identify the word type of the template word in the mining template and to identify the word relationship of the template word in the mining template;

[0210] Generate negative search terms from the template terms of the negative term type and the negative input terms;

[0211] Generate positive search terms from the template terms of the positive term type and the target terms;

[0212] The word relationships, negative search terms, positive search terms, mining relationships, and word distances are written into a preset query statement to obtain the search statement.

[0213] By identifying the word type, words of the negative type can be filtered from the template words, and negative search words can be generated comprehensively by combining the negative input words. By filtering words of the positive type from the template words, positive search words can be generated comprehensively by combining the positive input words. At the same time, the search statement can be generated quickly through the preset query statement, thereby improving the efficiency of text mining.

[0214] In at least one embodiment of the present invention, the generating unit 111 is further configured to count the number of searches of the search text;

[0215] A search ratio is generated based on the ratio of the number of searches to the number of texts.

[0216] Count the frequency of each positive search term in the search text;

[0217] The matching degree between the search text and the search statement is calculated based on the frequency.

[0218] Based on the matching degree, the target text is selected from the multiple search texts;

[0219] The number of texts, the number of searches, the search ratio, the frequency, and the target text are written into a preset interface to obtain the mining details interface.

[0220] By statistically analyzing the frequency of each positive search term, the matching degree can be accurately quantified. The matching degree can then be used to reasonably filter the target text, which is then displayed in the mining details interface, improving the intuitiveness of the target text.

[0221] In at least one embodiment of the present invention, the execution unit 113 is further configured to quantify the public opinion text stored in the preset microservice architecture based on the template type to obtain the text quantity;

[0222] Obtain the processing rate of idle threads from the preset search engine;

[0223] Extract the mining time requirement from the text mining request;

[0224] The processing rate with the smallest value is determined as the target rate;

[0225] The number of threads is calculated based on the number of texts, the required mining time, and the target rate.

[0226] The search thread is selected from the plurality of idle threads based on the number of threads and the processing rate;

[0227] The search thread is invoked to compare the search statement with the public opinion text, and the public opinion text that successfully matches the search statement is identified as the search text.

[0228] In this embodiment, since the text mining request is a further limited search on the mining template, the number of texts can be reasonably quantified by the template type. Then, by selecting the processing rate with the smallest value as the target rate, the number of threads can be determined as large as possible. Thus, by combining the number of threads with the processing rate, the search threads can be filtered out to ensure that the text mining request can be responded to within the mining time requirement.

[0229] As can be seen from the above technical solutions, this application, by identifying the mining template corresponding to the text mining request, can quickly generate the search statement and improve the efficiency of search text generation because the user does not need to repeatedly input the template content in the operation interface. When the mining template is detected to be triggered, it indicates that the mining template needs further improvement. Furthermore, by using the positive input words, it can generate associated words to assist the user's input, thereby improving the efficiency of target word generation. Simultaneously, it can assist in locking the mining scope when the user lacks structured tags. Combining the template words, the negative input words, the mining relationships, and the word distances, it can accurately generate search statements, thereby improving the accuracy of search text generation. By using the number of texts, it can reasonably call the search thread to process the search statement, thus achieving rapid processing of the search text without affecting other request processing. In addition, this application also uses the number of texts and the search text to generate a mining details interface, which can help the user quickly understand the response status of the text mining request.

[0230] like Figure 4 The diagram shown is a schematic diagram of the structure of an electronic device that implements the text mining method of the present invention.

[0231] In one embodiment of the present invention, the electronic device 1 includes, but is not limited to, a memory 12, a processor 13, and computer-readable instructions, such as a text mining program, stored in the memory 12 and executable on the processor 13.

[0232] Those skilled in the art will understand that the schematic diagram is merely an example of electronic device 1 and does not constitute a limitation on electronic device 1. It may include more or fewer components than shown in the diagram, or combine certain components, or different components. For example, electronic device 1 may also include input / output devices, network access devices, buses, etc.

[0233] The processor 13 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor 13 is the computing core and control center of the electronic device 1, connecting various parts of the electronic device 1 through various interfaces and lines, and executing the operating system of the electronic device 1, as well as various installed application programs and program code.

[0234] For example, the computer-readable instructions can be divided into one or more modules / units, which are stored in the memory 12 and executed by the processor 13 to complete the present invention. The one or more modules / units can be a series of computer-readable instruction segments capable of performing a specific function, which describe the execution process of the computer-readable instructions in the electronic device 1. For example, the computer-readable instructions can be divided into a display unit 110, a generation unit 111, an identification unit 112, an execution unit 113, a determination unit 114, an acquisition unit 115, and a detection unit 116.

[0235] The memory 12 can be used to store the computer-readable instructions and / or modules. The processor 13 implements various functions of the electronic device 1 by running or executing the computer-readable instructions and / or modules stored in the memory 12 and calling the data stored in the memory 12. The memory 12 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. The memory 12 may include non-volatile and volatile memory, such as: hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other storage devices.

[0236] The memory 12 can be the external memory and / or internal memory of the electronic device 1. Furthermore, the memory 12 can be a physical memory, such as a memory module, a TF card (Trans-flash Card), etc.

[0237] If the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by instructing related hardware through computer-readable instructions. The computer-readable instructions can be stored in a computer-readable storage medium, and when executed by a processor, the computer-readable instructions can implement the steps of the various method embodiments described above.

[0238] The computer-readable instructions include computer-readable instruction code, which may be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer-readable instruction code, recording medium, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), and random access memory (RAM).

[0239] The blockchain referred to in this invention is a novel application model of computer technologies such as distributed text mining, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0240] Combination Figure 1 The memory 12 in the electronic device 1 stores computer-readable instructions to implement a text mining method, and the processor 13 can execute the computer-readable instructions to achieve the following:

[0241] When a text mining request is received, a mining template is displayed in the operation interface according to the text mining request;

[0242] When it is detected that the mining template is successfully triggered in the operation interface, target words are generated based on the positive input words in the operation interface. The target words include the positive input words and the keywords in the associated words corresponding to the positive input words.

[0243] The mining relationship is identified based on the input position of the target words in the operation interface;

[0244] A search statement is generated based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words;

[0245] Based on the number of texts in the preset microservice architecture corresponding to the template type of the mining template, the search thread in the preset search engine is invoked to execute the search statement and obtain the search text;

[0246] A mining details interface is generated based on the number of texts and the search text.

[0247] Specifically, the specific implementation method of the processor 13 for the above-mentioned computer-readable instructions can be found in [reference needed]. Figure 1 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0248] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0249] The computer-readable storage medium stores computer-readable instructions, which, when executed by the processor 13, are used to perform the following steps:

[0250] When a text mining request is received, a mining template is displayed in the operation interface according to the text mining request;

[0251] When it is detected that the mining template is successfully triggered in the operation interface, target words are generated based on the positive input words in the operation interface. The target words include the positive input words and the keywords in the associated words corresponding to the positive input words.

[0252] The mining relationship is identified based on the input position of the target words in the operation interface;

[0253] A search statement is generated based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words;

[0254] Based on the number of texts in the preset microservice architecture corresponding to the template type of the mining template, the search thread in the preset search engine is invoked to execute the search statement and obtain the search text;

[0255] A mining details interface is generated based on the number of texts and the search text.

[0256] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0257] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0258] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0259] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units or devices described may also be implemented by a single unit or device through software or hardware. Terms such as "first," "second," etc., are used to indicate names and do not indicate any specific order.

[0260] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A text mining method, characterized in that, The text mining method includes: When a text mining request is received, a mining template is displayed in the operation interface according to the text mining request. The text box position in the operation interface includes positive and negative positions. The mining template refers to a template that matches the scene type corresponding to the request scene. The method of determining the mining template includes: parsing the request message of the text mining request to obtain the data information carried by the request message; extracting the request scene from the data information; identifying the scene type corresponding to the request scene as the template type; and obtaining a template corresponding to the template type from a preset template library as the mining template. When the mining template is successfully triggered in the operation interface, target words are generated based on the positive input words in the operation interface, including: obtaining the positive input words from the positive position; generating associated words corresponding to the positive input words based on a pre-trained synonym prediction model; detecting whether the associated words are successfully triggered; determining the successfully triggered associated words as keywords; and generating the target words based on the positive input words and the keywords, wherein the target words include the positive input words and the keywords, and the positive input words refer to words that must be included in the search text. The mining relationship is identified based on the input position of the target words in the operation interface; A search statement is generated based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words; Based on the number of texts of the template type corresponding to the mining template in the preset microservice architecture, the search thread in the preset search engine is invoked to execute the search statement to obtain the search text; A mining details interface is generated based on the number of texts and the search text.

2. The text mining method as described in claim 1, characterized in that, Before generating target words based on the positive input words in the user interface, the text mining method further includes: Identify the initial display time of the excavation template on the operation interface; Based on the initial display time, obtain the operation log of the operation interface; Identify the button positions of the associated buttons related to the mining template from the operation interface, and identify the box positions of the text boxes in the operation interface; Check whether the operation log includes the button position and the box position; If the operation log includes the button position and / or the box position, then it is determined that the mining template was successfully triggered in the operation interface.

3. The text mining method as described in claim 1, characterized in that, The synonym prediction model includes a semantic analysis network and a prediction output network. The semantic analysis network includes a forward feature extraction network and a backward feature extraction network. The synonym prediction model, based on pre-trained data, generates associated words corresponding to the positive input words, including: The positive input vocabulary is encoded to obtain the input vector; Based on the positive feature extraction network, features are extracted from the input vector to obtain the first feature; Based on the inverse feature extraction network, features are extracted from the input vector to obtain the second feature; Generate semantic vectors for the positive input words based on the first feature and the second feature; Based on the semantic vector, the corresponding words are obtained from the prediction output network as the associated words.

4. The text mining method as described in claim 1, characterized in that, The step of generating a search statement based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words includes: Identify the word type of the template words in the mining template, and identify the word relationships of the template words in the mining template; Generate negative search terms from the template terms of the negative term type and the negative input terms; Generate positive search terms from the template terms of the positive term type and the target terms; The word relationships, negative search terms, positive search terms, mining relationships, and word distances are written into a preset query statement to obtain the search statement.

5. The text mining method as described in claim 4, characterized in that, The step of generating a mining details interface based on the number of texts and the search text includes: Count the number of searches for the search text; A search ratio is generated based on the ratio of the number of searches to the number of texts. Count the frequency of each positive search term in the search text; The matching degree between the search text and the search statement is calculated based on the frequency. Based on the matching degree, the target text is selected from the multiple search texts; The number of texts, the number of searches, the search ratio, the frequency, and the target text are written into a preset interface to obtain the mining details interface.

6. The text mining method as described in claim 1, characterized in that, The step involves calling a search thread in a preset search engine to execute the search statement based on the number of texts of the template type corresponding to the mining template within a preset microservice architecture, resulting in the following search texts: Based on the template type, the public opinion texts stored in the preset microservice architecture are quantified to obtain the number of texts; Obtain the processing rate of idle threads from the preset search engine; Extract the mining duration requirement from the text mining request; The processing rate with the smallest value is determined as the target rate; The number of threads is calculated based on the number of texts, the required mining time, and the target rate. The search thread is selected from the plurality of idle threads based on the number of threads and the processing rate; The search thread is invoked to compare the search statement with the public opinion text, and the public opinion text that successfully matches the search statement is identified as the search text.

7. A text mining device, characterized in that, The text mining device includes: The display unit is used to display a mining template in the operation interface when a text mining request is received, according to the text mining request. The text box position in the operation interface includes positive and negative positions. The mining template refers to a template that matches the scene type corresponding to the request scene. The method of determining the mining template includes: parsing the request message of the text mining request to obtain the data information carried by the request message; extracting the request scene from the data information; identifying the scene type corresponding to the request scene as the template type; and obtaining a template corresponding to the template type from a preset template library as the mining template. A generation unit is configured to generate target words based on positive input words in the operation interface when the mining template is successfully triggered in the operation interface. This includes: obtaining the positive input words from the positive position; generating associated words corresponding to the positive input words based on a pre-trained synonym prediction model; detecting whether the associated words are successfully triggered; identifying the successfully triggered associated words as keywords; and generating the target words based on the positive input words and the keywords. The target words include the positive input words and the keywords, where the positive input words refer to words that must be included in the search text. The identification unit is used to identify the mining relationship based on the input position of the target word in the operation interface; The generation unit is further configured to generate search statements based on the template words in the mining template, the negative input words in the operation interface, the mining relationship, and the word distance between the target words; An execution unit is used to execute the search statement by calling a search thread in a preset search engine based on the number of texts of the template type corresponding to the mining template in a preset microservice architecture, so as to obtain the search text; The generation unit is also used to generate a mining details interface based on the number of texts and the search text.

8. An electronic device, characterized in that, The electronic device includes: Memory, which stores computer-readable instructions; and The processor executes computer-readable instructions stored in the memory to implement the text mining method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions that are executed by a processor in an electronic device to implement the text mining method as described in any one of claims 1 to 6.

Images