CN115080720A - Text processing method, device, equipment and medium based on RPA and AI - Google Patents
Text processing method, device, equipment and medium based on RPA and AI Download PDFInfo
- Publication number
- CN115080720A CN115080720A CN202210750317.1A CN202210750317A CN115080720A CN 115080720 A CN115080720 A CN 115080720A CN 202210750317 A CN202210750317 A CN 202210750317A CN 115080720 A CN115080720 A CN 115080720A
- Authority
- CN
- China
- Prior art keywords
- text
- frequency
- question
- information
- answer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 51
- 230000008451 emotion Effects 0.000 claims abstract description 36
- 230000008569 process Effects 0.000 claims abstract description 36
- 238000012545 processing Methods 0.000 claims abstract description 31
- 238000004458 analytical method Methods 0.000 claims abstract description 20
- 230000003993 interaction Effects 0.000 claims abstract description 17
- 238000000605 extraction Methods 0.000 claims abstract description 12
- 238000012163 sequencing technique Methods 0.000 claims description 9
- 230000002996 emotional effect Effects 0.000 claims description 8
- 238000012216 screening Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 16
- 230000006854 communication Effects 0.000 description 11
- 238000004891 communication Methods 0.000 description 10
- BUGBHKTXTAQXES-UHFFFAOYSA-N Selenium Chemical compound [Se] BUGBHKTXTAQXES-UHFFFAOYSA-N 0.000 description 5
- 229910052711 selenium Inorganic materials 0.000 description 5
- 239000011669 selenium Substances 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000004801 process automation Methods 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- VYZAMTAEIAYCRO-UHFFFAOYSA-N Chromium Chemical compound [Cr] VYZAMTAEIAYCRO-UHFFFAOYSA-N 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000003203 everyday effect Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000003032 molecular docking Methods 0.000 description 2
- 238000004806 packaging method and process Methods 0.000 description 2
- 230000003252 repetitive effect Effects 0.000 description 2
- 208000025174 PANDAS Diseases 0.000 description 1
- 208000021155 Paediatric autoimmune neuropsychiatric disorders associated with streptococcal infection Diseases 0.000 description 1
- 240000000220 Panda oleosa Species 0.000 description 1
- 235000016496 Panda oleosa Nutrition 0.000 description 1
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005111 flow chemistry technique Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Human Computer Interaction (AREA)
- Machine Translation (AREA)
Abstract
The application discloses a text processing method, a text processing device, text processing equipment and a text processing medium based on RPA and AI, wherein the text processing method comprises the following steps: obtaining chatting detail information generated in the interaction process of e-commerce customer service and customers of each virtual shop in a seller workbench application, wherein the chatting detail information comprises a plurality of questions and original information texts corresponding to the questions; performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions; performing information extraction and text emotion analysis on the original information text of each high-frequency question to obtain an answer of the high-frequency question; combining each high-frequency question and the answer of the high-frequency question into a target text; all the target texts are written into a background QA knowledge base, so that a large amount of manpower is not required to be concentrated on message monitoring and collection, and the high-frequency problems and answers thereof are not required to be rechecked and configured manually, so that the processing efficiency can be improved, and the fine management of merchants is facilitated.
Description
Technical Field
The present application relates to the field of process automation technologies, and in particular, to a method, an apparatus, a device, and a medium for processing a text based on an RPA.
Background
With the rapid development of society, the e-commerce industry plays an increasingly important role. At present, a plurality of bottom layer frames are used for secondary building and packaging in the market to construct a set of automatic work flow system, and in the work flow system, a flow automatic robot is mainly used for flow processing. A Process Automation Robot (RPA) is a product solution that automates manual activities by performing repetitive rule-based tasks. RPA is mainly focused on using a human-machine interface to perform automated processing instead of manual operation.
However, in this case, the merchant needs to manually configure all store robot basic docking and integration information once, then manually collect and gather chat data of each store every day to form common question answers (FAQ), and finally manually enter the FAQ big data into the corresponding software robot system. The FAQ big data needs to be collected, configured and checked manually, the online self-retrieval of chatting data cannot be realized, the chatting data can not be retrieved and pulled and deposited in a robot background database, high-frequency problems and answers need to be rechecked and configured manually, and the processing efficiency is reduced.
Disclosure of Invention
In view of the foregoing, the present application is proposed to provide a text processing method, apparatus, device, and medium based on RPA and AI. The specific scheme is as follows:
a text processing method based on RPA and AI includes:
obtaining chatting detail information generated in the interaction process of e-commerce customer service and customers of each virtual shop in a seller workbench application, wherein the chatting detail information comprises a plurality of questions and original information texts corresponding to the questions;
performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions;
performing information extraction and text emotion analysis on the original information text of each high-frequency question to obtain an answer of the high-frequency question;
combining each high-frequency question and the answer of the high-frequency question into a target text;
and writing all the target texts into a background QA knowledge base.
Preferably, the performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions comprises:
identifying a named entity in each question in the chat detail information to obtain all vocabularies in each question;
performing word frequency statistics on all the words in each question to obtain the word frequency of all the words in each question;
sequencing the word frequencies of all the vocabularies to obtain a sequencing result;
determining a plurality of high-frequency words according to the sequencing result;
and taking the question containing the high-frequency vocabulary in each question as a high-frequency question.
Preferably, the extracting information and analyzing text emotion for the original information text of each high-frequency question to obtain an answer to the high-frequency question includes:
performing emotion attribute identification on the original information text of each high-frequency question to obtain the emotion attribute of the original information text of each high-frequency question;
screening each original information text according to the emotional attribute of each original information text to obtain an intermediate information text of each high-frequency problem;
performing semantic recognition on the intermediate information text of each high-frequency problem to obtain semantic information of the intermediate information text of each high-frequency problem;
matching the semantic information of the intermediate information text of each high-frequency question with the semantic information of each standard answer in a preset standard answer library to obtain a matching result;
and determining answers of the high-frequency questions based on the matching results.
Preferably, the determining answers to the high-frequency questions based on the matching results includes:
and according to the matching result, taking the intermediate information text matched with the standard answer in the intermediate information text of each high-frequency question as the answer of the corresponding high-frequency question, removing the intermediate information text which is not matched with the standard answer in the intermediate information text of each high-frequency question, and taking the standard answer corresponding to the high-frequency question with the removed intermediate information text as the answer.
Preferably, the obtaining chat detail information generated in the interaction process between the e-commerce customer service of each virtual store in the vendor workbench application and the customer includes:
starting a seller workbench application window;
entering a chat window interface of each virtual store;
and acquiring webpage data on the chat window interface of each shop, and taking the acquired webpage data as chat detail information.
Preferably, the collecting the webpage data on the chat window interface of each store, and using the collected webpage data as chat detail information includes:
acquiring all target links of a current webpage;
clicking all the target links, and entering a detail page corresponding to each target link;
extracting data in a detail page corresponding to each target link, and taking the extracted data in each detail page as chatting detail data;
and searching whether a page turning button exists in the current webpage or not, if so, clicking the page turning button to enter the next webpage, taking the next webpage as the current webpage, and returning to the step of acquiring all target links of the current webpage until the page turning button cannot be searched.
Preferably, the launching of the seller workbench application window comprises:
starting a browser driver using an automated testing tool to start a vendor workstation application;
and starting an application window of the seller workbench application through remote driving.
A text processing device based on RPA and AI includes:
the information acquisition unit is used for acquiring chatting detail information of each shop in a seller workbench application, wherein the chatting detail information is generated in the interaction process of an e-commerce customer service and a customer, and the chatting detail information of each shop comprises a plurality of questions and original information texts corresponding to the questions;
the high-frequency question acquisition unit is used for performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions;
the answer obtaining unit is used for extracting information and performing text emotion analysis on the original information text of each high-frequency question to obtain an answer of the high-frequency question;
the text generation unit is used for combining each high-frequency question and the answer of the high-frequency question into a target text;
and the text writing unit is used for writing all the target texts into a background QA knowledge base.
A RPA and AI based text processing apparatus comprising: a memory and a processor;
the memory is used for storing programs;
the processor is configured to execute the program to implement the steps of the RPA and AI based text processing method.
A storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the RPA and AI based text processing method as described above.
By means of the technical scheme, the text processing method based on the RPA and the AI obtains chat detail information generated in the interaction process of the E-commerce customer service and the customer of each virtual shop in the application of a workbench of a seller, performs text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions, performs information extraction and text emotion analysis on the original information text of each high-frequency question to obtain the answer of the high-frequency question, combines each high-frequency question and the answer of the high-frequency question into one target text, and finally writes all the target texts into a background QA knowledge base. The high-frequency questions and answers thereof do not need to be rechecked and configured manually any more, so that the processing efficiency can be improved, and the fine management of merchants can be realized.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the application. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
fig. 1 is a schematic flowchart of a text processing method based on RPA and AI according to an embodiment of the present disclosure;
fig. 2 is a schematic structural diagram of a text processing apparatus based on RPA and AI according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of a text processing device based on RPA and AI according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Under the current working situation of the e-commerce, a merchant needs to manually configure all store robot foundation docking and integration information once, then manually collect and gather chatting data of each store every day to form common problem answers (FAQ), and finally manually input FAQ big data into a corresponding software robot system. The FAQ big data needs to be collected, configured and checked manually, the online self-retrieval of chatting data cannot be realized, the chatting data can not be retrieved and pulled and deposited in a robot background database, high-frequency problems and answers need to be rechecked and configured manually, and the processing efficiency is reduced.
Therefore, in order to solve the above problems, the present application provides a text processing method, device, apparatus and medium based on RPA and AI, which can determine a high frequency question and an answer corresponding to the high frequency question, thereby avoiding manually reviewing and configuring the high frequency question and the answer. By fusing the RPA and the AI, the method can quickly form a special requirement scene in a front-end workflow manner, and meanwhile, the workflow is configured for one time in a plurality of shops, so that the instant use and automatic monitoring can be realized through bottom-layer drive communication in the subsequent use process; the system has the advantages that the emotional attribute and semantic information are locked through modeling by means of OCR and NLP, data grabbing, auditing and database falling are completed in a real-time traversing cycle mode, efficient, rapid and automatic replying is achieved, the language processing and problem type positioning functions are achieved, the problem that frequently answers cannot be achieved is avoided, the processing efficiency can be improved, and the system helps merchants to achieve fine management and digital operation.
Referring to fig. 1, fig. 1 is a schematic flow chart of a text processing method based on RPA and AI provided in an embodiment of the present application, where the method includes:
step S110, obtaining chatting detail information generated in the interaction process of the e-commerce customer service and the customer of each virtual shop in the application of a workbench of a seller, wherein the chatting detail information comprises a plurality of questions and original information texts corresponding to the questions.
Particularly, the seller workbench is a platform for integrating store management tools, operation and consultation information and business partnership for e-commerce sellers, and can improve the processing efficiency of the sellers on events. The chatting detail information can be history chatting information records generated in the interaction process of e-commerce customer service and customers of each virtual shop in the application of a workbench of a seller, and the e-commerce customer service can be a real person or a robot. The customer can initiate a question about the information of the commodity, the e-commerce customer service communicates with the customer according to the question of the customer, and knows the core requirement of the customer according to the communication process, so as to solve the question of the customer, so that the chat detail information can comprise a plurality of questions proposed by different customers, and a long-section dialogue which is developed around the question, the dialogue comprises an intermediate process field generated by the communication and an answer field finally used for solving, and the intermediate process field and the answer field are both included in an original information text corresponding to each question.
Further, in this embodiment, the chat-related information is acquired by loading the acquisition function module in the RPA robot. RPA, a Process Automation robot (Robotic Process Automation), is a product solution for automating manual activities by performing repetitive rule-based tasks, and has the advantages of high efficiency, low error rate, reduced labor, etc.
And step S120, performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions.
Specifically, in this step, a high-frequency problem is determined by running a language-text recognition processing function module in the RPA robot. Language text recognition, which is capable of recognizing language text from pictures, has been modularized into RPA robots as one of the functions that RPA robots can implement. The high frequency problem is determined from the picture by loading the corresponding module of the RPA robot, thereby using language text recognition. In some embodiments, the chat detail information may be stored as a file in a picture format, so that a language text recognition technology is applied to the picture to determine the high frequency question, and in some other embodiments, the chat detail information may be stored as a file in another format, and then the high frequency question is recognized by using another text recognition technology corresponding to the format, so that the high frequency question is determined.
And step S130, performing information extraction and text emotion analysis on the original information text of each high-frequency question to obtain an answer of the high-frequency question.
Specifically, after the chat detail information is acquired, the chat details of the e-commerce customer service and the customer are obtained, wherein the chat details comprise questions of the customer, answers made by the customer service and useless fields, and words or pictures which are not obviously related to the questions and the answers can be included. And determining an answer corresponding to each high-frequency question by operating an information extraction and text emotion analysis functional module in the RPA robot. Information extraction and text emotion analysis have been modularized into RPA robots as one of the functions that RPA robots can implement. And (4) performing information extraction and text emotion analysis on the original information text by loading a corresponding module of the RPA robot so as to determine the answer of the high-frequency question.
And step S140, combining each high-frequency question and the answer of the high-frequency question into a target text.
Specifically, after the high-frequency question and the answer corresponding to the high-frequency question are determined, the high-frequency question and the answer corresponding to the high-frequency question can be combined into a target text.
It can be understood that when the shop sets an unattended mode, an automatic mechanism of the software robot is triggered, after a user proposes a relevant question, the software robot performs big data traversal on the background at the moment, when the relevant question and a corresponding answer are traversed, the software robot can automatically reply, if the relevant question and the corresponding answer are not traversed, the answer corresponding to the question cannot be obtained, and at the moment, the software robot is silent and cannot answer the question of the customer. In this embodiment, when the problem posed by the client is faced, the software robot may determine the answer directly according to the problem posed by the client, and the problem of silence without traversing is avoided.
And S150, writing all the target texts into a background QA knowledge base.
Specifically, in the present embodiment, all target texts are written into the background QA knowledge base by loading the traversal card of the RPA robot. When the problem brought forward by the client is faced, the software robot can directly search for similar problems in a background QA knowledge base according to the problem brought forward by the client in an ergodic way, then the answer is determined according to the problems, and the problem that the software robot is silent when the software robot cannot be traversed is avoided.
From the technical scheme, the text processing method based on the RPA and the AI obtains chatting detail information generated in the interaction process of the E-commerce customer service and the customer of each virtual shop in the application of a workbench of a seller, performs text recognition and data statistics on each question in the chatting detail information to obtain a plurality of high-frequency questions, performs information extraction and text emotion analysis on the original information text of each high-frequency question to obtain the answer of the high-frequency question, combines each high-frequency question and the answer of the high-frequency question into one target text, and writes all the target texts into a background QA knowledge base, wherein after the problems and the original information text corresponding to each problem are determined, the high-frequency question is determined, then the answer corresponding to the high-frequency question is determined, so that a large amount of manpower does not need to be concentrated on message monitoring and collection, the high-frequency questions and answers thereof do not need to be rechecked and configured manually any more, so that the processing efficiency can be improved, and the fine management of merchants can be realized.
The above embodiments briefly introduce a text processing method based on RPA and AI in the present application. In some embodiments of the present application, for the step S120, performing text recognition and data statistics on each question in the chat detail information to obtain a detailed description of a process of obtaining a plurality of high-frequency questions, where the process may include the following steps:
step S121, identifying the named entity in each question in the chat detail information to obtain all vocabularies in each question.
Specifically, after the questions are subjected to text arrangement, all words in each question are extracted in a LAC named entity recognition mode. Lac (lexical Analysis of chinese) is a joint lexical Analysis tool, and can implement functions of chinese word segmentation, part of speech tagging, proper name recognition, and the like. For example, enter a sentence: "LAC is an excellent word segmentation tool" and word segmentation yields the result: "LAC", "is", "individual", "excellent", "of", "word segmentation", "tool".
Further, in this embodiment, the organizational entity in the text message can be extracted by calling the Taskflow API in PaddleNLP to get all the words in each question.
And step S122, performing word frequency statistics on all the words in each question to obtain the word frequency of all the words in each question.
Specifically, word frequency statistics are performed on all the words in all the questions to determine the word frequency of all the words in each question. It should be noted that the word frequency statistics herein is for all words in all questions, and it can be seen from the word frequency statistics that those particular words are high frequency words. Furthermore, the specific distribution of a certain high-frequency vocabulary or certain high-frequency vocabularies under certain problems or certain specific scenes can be known, and the high-frequency problems can be determined.
And S123, sequencing the word frequencies of all the vocabularies to obtain a sequencing result.
Specifically, the words of all the words are sorted from high to low or from low to high according to their word frequencies, so that the frequency of which word or words appear in all the questions is known to be high. In some embodiments, according to the sorting result, the average word frequency of all the words may be calculated as a criterion for measuring the high-frequency words, or a value may be preset, and a word whose word frequency is higher than the preset value may be determined as a high-frequency word.
And step S124, determining a plurality of high-frequency vocabularies according to the sequencing result.
Specifically, after the sorting result is obtained, the average word frequency of all words can be calculated, which is used as a standard for measuring high-frequency words, or a value can be preset, and words whose word frequency is higher than the preset value are used as high-frequency words, which may include some less important query words, inflectives, and query punctuation marks (such as "why", "do", and ".
And step S125, regarding the question containing the high-frequency vocabulary in each question as a high-frequency question.
In particular, the high frequency problem herein may include an unlimited number and unlimited frequency of words of high frequency vocabulary.
Further, for example, customers often compare a product to its replacement products, and similar problems arise as follows: "what is the product's performance related to price? The question may include at least two high frequency words, performance and price, belonging to the high frequency question.
In the above embodiment, the process of performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions is briefly introduced. In some embodiments of the present application, for step S130, a process of extracting information and performing text sentiment analysis on the original information text of each high-frequency question to obtain an answer to the high-frequency question is described in detail, where the process may include the following steps:
step S131, performing emotion attribute identification on the original information text of each high-frequency question to obtain the emotion attribute of the original information text of each high-frequency question.
Specifically, in the step, the original information text of each high-frequency problem is segmented and cut, the original information text is divided into a plurality of sections of sentences, all phrases in the section of sentences are extracted from each section of words, then the phrases are subjected to emotion attribute recognition, the emotion attributes of the phrases are obtained, so that the emotion attributes of the section of sentences are determined, the hierarchical phenomenon of emotion characteristics and the relation among characteristics are analyzed according to typical emotion characteristic algorithm models of all layers of classification levels, and then the emotion attributes are classified in an algorithm bottom layer system.
Furthermore, the emotion attribute recognition is one of the functions of language text recognition, and the emotion attribute recognition of the sentence is helpful for accurate analysis and positioning of answers. In the emotion attribute recognition, generally, the emotion of a sentence or a phrase can be classified into positive, negative, neutral, positive, negative, and the like, and by means of the emotion attribute analysis of the sentence or the phrase, it can be preliminarily determined under what scenario the user asks, what question the user generally asks, what information the user wants to obtain through the question, and the like. The emotion attribute identification is beneficial to determining answers of high-frequency questions preliminarily, and lays a foundation for further range reduction subsequently.
And S132, screening each original information text according to the emotional attribute of each original information text to obtain the intermediate information text of each high-frequency problem.
Specifically, the original message text not only includes answers to the high-frequency questions, but also includes some intermediate procedures for communication, animation expressions, pictures, and the like. If the customer service reply message does not completely solve the user's problem, the user may continue to ask the question with questioning or negative emotional attributes, and the customer service message may be determined not to be the answer to the question, and the customer service message may be screened out.
Step S133, performing semantic identification on the intermediate information text of each high-frequency question to obtain semantic information of the intermediate information text of each high-frequency question.
Specifically, after the intermediate information text of each high-frequency problem is obtained through emotion attribute identification and screening, semantic identification is performed on the intermediate information text, so that the meaning of the intermediate information text, namely semantic information, is known, and the next operation is performed.
And S134, matching the semantic information of the intermediate information text of each high-frequency question with the semantic information of each standard answer in a preset standard answer library to obtain a matching result.
Specifically, semantic information of each standard answer in the preset answer library is acquired, and semantic information of the intermediate information text of each high-frequency question is matched with semantic information of each standard answer in the preset standard answer library, so that whether the semantic information of the intermediate text of the high-frequency question meets the requirement of the standard answer can be determined, and whether the intermediate information text is the answer of the high-frequency question or not can be determined.
Furthermore, in a specific use scene, an enterprise only needs to continuously optimize the standard answer library without needing a customer service specialist to concentrate on monitoring and collecting enterprise messages, so that the labor is saved, the problem of simultaneous online and offline operation configuration of multiple shops can be easily solved, the effect of noninductive use is achieved, and the enterprise is really helped to realize refined operation.
And step S135, determining answers of the high-frequency questions based on the matching results.
Specifically, if the semantic information of the intermediate text of the high-frequency question meets the standard answer requirement, the intermediate information text may be determined as the answer of the high-frequency question.
In the above embodiment, the process of extracting information and performing text emotion analysis on the original information text of each high-frequency question to obtain the answer to the high-frequency question is simply introduced. In some embodiments of the present application, the detailed description of the process of determining the answer to each of the high-frequency questions based on the matching result in step S135 may include the following steps:
and according to the matching result, taking the intermediate information text matched with the standard answer in the intermediate information text of each high-frequency question as the answer of the corresponding high-frequency question, removing the intermediate information text which is not matched with the standard answer in the intermediate information text of each high-frequency question, and taking the standard answer corresponding to the high-frequency question with the removed intermediate information text as the answer.
Specifically, an intermediate information text with certain emotional attributes and semantic information thereof in accordance with standard answers is used as an answer of the high-frequency question, however, if after emotional attribute recognition, screening, semantic recognition and matching are carried out, the intermediate information text is found to have no related emotional attributes, and the semantic information thereof is not matched with the semantic information of the standard answers, the problem proposed by a client is not solved by the intermediate text information, the intermediate information text is rejected as useless text information, and the standard answer corresponding to the high-frequency question from which the intermediate information text is rejected is used as the answer of the high-frequency question.
The above-mentioned embodiments briefly describe the process of determining the answer to each of the high-frequency questions based on the matching result. In some embodiments of the present application, for step S110, a detailed description is performed on a process of obtaining chat detail information generated in an interaction process between an e-commerce customer service and a customer of each virtual store in the vendor workstation application, where the process may include the following steps:
and step S111, starting the application window of the workbench of the seller.
Specifically, in this step, the seller workbench is opened by loading the application starting function module in the RPA robot. The vendor workstation may be a knurl net. The specific operation steps are as follows:
chromedriver is launched by the underlying Selenium to open the vendor's workbench on the browser interface.
The Selenium is an automatic testing tool and is used for driving a browser to execute corresponding operations, and the Chrome driver is Chrome driver.
And starting a browser window by a remote driver to open an application window of a workbench of the seller.
Specifically, remote driver is a remote drive.
Further, starting Chromedriver through the underlying Selenium to open the vendor workbench on the browser interface may include:
and inputting and accessing the URL to call an Abstract HttpCommandec method in the Selenium, requesting a driver to process, finally, interacting the driver with a browser, and making corresponding operation by the browser according to the instruction.
Specifically, entering and accessing a URL is by entering a uniform resource locator to access the uniform resource locator system and invoking an HTTP commander in the Selenium.
The browser driver is started using HTTP command symbols.
And starting the seller workbench application by using the browser driver.
And step S112, entering a chat window interface of each virtual shop.
Specifically, in this step, the interface element click function module in the RPA robot is loaded, and a corresponding event on the browser interface is clicked, so as to enter a chat window of each shop.
And S113, acquiring webpage data on a chat window interface of each shop, and taking the acquired webpage data as chat detail information.
Specifically, in this step, coordinate setting and automatic page turning polling traversal are performed through an intelligent web page data acquisition module of the RPA robot, so as to obtain chat detail data.
Furthermore, the steps effectively avoid repeated acquisition of node data, keep real-time incremental updating and acquisition of data, reduce the program occupancy rate and improve the processing efficiency.
The above embodiment provides a simple introduction to the process of acquiring chat detail information generated in the interaction process between the e-commerce customer service and the customer of each virtual shop in the application of the vendor workbench. In some embodiments of the application, for step S113, the web page data on the chat window interface of each store is collected, and the collected web page data is used as a detailed introduction to the process of chat detail information, where the process may include the following steps:
and step S111, acquiring all target links of the current webpage.
And S112, clicking all the target links, and entering a detail page corresponding to each target link.
And step S113, extracting data in the detail page corresponding to each target link, and taking the extracted data in each detail page as chatting detail data.
And step S114, searching whether a page turning button exists in the current webpage, if so, clicking the page turning button to enter the next webpage, taking the next webpage as the current webpage, and returning to the step of acquiring all target links of the current webpage until the page turning button cannot be searched.
Specifically, through the steps, chat detail data generated by interaction of all E-commerce customer services and customers can be collected.
Further, after all chatting detail data are collected, the chatting detail data are stored into a file with a certain format. In some embodiments, chat detail data is obtained to a local form, the local form is converted to a local picture, and then an OCR and NLP language text processing function in the RPA robot is run to obtain a plurality of questions and original information text corresponding to each of the questions. In some other embodiments, the chatting detail data may be stored as a file in other format, and then a plurality of questions and original information text corresponding to each of the questions are obtained in the file.
The above process, which may include:
and step S1, acquiring the chatting detail data to a local table.
Specifically, the step may include:
and calling a class method in the bottom-layer pandas library to read Excel from the acquisition result.
And constructing a dataframe by using a dictionary, importing the data in the Excel to initialize the dataframe, and obtaining a local table.
Specifically, the dictionary is a variable container model and can store any type of object, specifically, in this embodiment, the dictionary is used to provide an index relationship between Excel and dataframe, and a correspondence relationship between data between Excel and dataframe is constructed by different Key values (Key-values) of the dictionary.
And removing the original index column of the local table.
Storing the local table.
And step S2, converting the local form into a local picture.
And step S3, operating an OCR and NLP language text processing function module in the RPA robot, thereby acquiring a plurality of questions and original information texts corresponding to the questions.
Specifically, the OCR and NLP bottom layers are fused into development packaging, so that the integrated functions of text image recognition, interception and annotation can be realized, and the efficiency is high.
The following describes the RPA and AI based text processing apparatus provided in the embodiments of the present application, and the RPA and AI based text processing apparatus described below and the RPA and AI based text processing method described above may be referred to correspondingly.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a text processing apparatus based on RPA and AI according to an embodiment of the present disclosure.
As shown in fig. 2, the apparatus may include:
the information acquisition unit 11 is configured to acquire chat detail information of each store in a vendor workbench application, where the chat detail information is generated in a process of interaction between an e-commerce customer service and a customer, and the chat detail information of each store includes a plurality of questions and an original information text corresponding to each question;
a high-frequency question obtaining unit 12, configured to perform text recognition and data statistics on each question in the chat detail information to obtain multiple high-frequency questions;
the answer obtaining unit 13 is configured to perform information extraction and text emotion analysis on the original information text of each high-frequency question to obtain an answer to the high-frequency question;
a text generating unit 14 for combining each of the high-frequency questions and the answer to the high-frequency question into one target text;
and a text writing unit 15, configured to write all the target texts into a background QA knowledge base.
The text processing device based on the RPA and the AI provided by the embodiment of the application can be applied to text processing equipment based on the RPA and the AI, such as a terminal: mobile phones, computers, etc. Optionally, fig. 3 is a block diagram illustrating a hardware structure of the RPA and AI-based text processing device, and referring to fig. 3, the hardware structure of the RPA and AI-based text processing device may include: at least one processor 1, at least one communication interface 2, at least one memory 3 and at least one communication bus 4;
in the embodiment of the application, the number of the processor 1, the communication interface 2, the memory 3 and the communication bus 4 is at least one, and the processor 1, the communication interface 2 and the memory 3 complete mutual communication through the communication bus 4;
the processor 1 may be a central processing unit CPU, or an application Specific Integrated circuit asic, or one or more Integrated circuits configured to implement embodiments of the present invention, etc.;
the memory 3 may include a high-speed RAM memory, and may further include a non-volatile memory (non-volatile memory) or the like, such as at least one disk memory;
wherein the memory stores a program and the processor can call the program stored in the memory, the program for:
obtaining chatting detail information generated in the interaction process of e-commerce customer service and customers of each virtual shop in a seller workbench application, wherein the chatting detail information comprises a plurality of questions and original information texts corresponding to the questions;
performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions;
performing information extraction and text emotion analysis on the original information text of each high-frequency question to obtain an answer of the high-frequency question;
combining each high-frequency question and the answer of the high-frequency question into a target text;
and writing all the target texts into a background QA knowledge base.
Alternatively, the detailed function and the extended function of the program may be as described above.
Embodiments of the present application further provide a storage medium, where a program suitable for execution by a processor may be stored, where the program is configured to:
obtaining chatting detail information generated in the interaction process of e-commerce customer service and customers of each virtual shop in a seller workbench application, wherein the chatting detail information comprises a plurality of questions and original information texts corresponding to the questions;
performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions;
performing information extraction and text emotion analysis on the original information text of each high-frequency question to obtain an answer of the high-frequency question;
combining each high-frequency question and the answer of the high-frequency question into a target text;
and writing all the target texts into a background QA knowledge base.
Alternatively, the detailed function and the extended function of the program may refer to the above description.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
In the present specification, the embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, the embodiments may be combined as needed, and the same and similar parts may be referred to each other.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A text processing method based on RPA and AI is characterized by comprising the following steps:
obtaining chatting detail information generated in the interaction process of e-commerce customer service and customers of each virtual shop in a seller workbench application, wherein the chatting detail information comprises a plurality of questions and original information texts corresponding to the questions;
performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions;
performing information extraction and text emotion analysis on the original information text of each high-frequency question to obtain an answer of the high-frequency question;
combining each high-frequency question and the answer of the high-frequency question into a target text;
and writing all the target texts into a background QA knowledge base.
2. The method of claim 1, wherein performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high frequency questions comprises:
identifying a named entity in each question in the chat detail information to obtain all vocabularies in each question;
performing word frequency statistics on all the words in each question to obtain the word frequency of all the words in each question;
sequencing the word frequencies of all the vocabularies to obtain a sequencing result;
determining a plurality of high-frequency words according to the sequencing result;
and taking the question containing the high-frequency vocabulary in each question as a high-frequency question.
3. The method of claim 1, wherein the extracting information and analyzing text emotion of the original information text of each high-frequency question to obtain an answer to the high-frequency question comprises:
performing emotion attribute identification on the original information text of each high-frequency question to obtain the emotion attribute of the original information text of each high-frequency question;
screening each original information text according to the emotional attribute of each original information text to obtain an intermediate information text of each high-frequency problem;
performing semantic recognition on the intermediate information text of each high-frequency problem to obtain semantic information of the intermediate information text of each high-frequency problem;
matching the semantic information of the intermediate information text of each high-frequency question with the semantic information of each standard answer in a preset standard answer library to obtain a matching result;
and determining answers of the high-frequency questions based on the matching results.
4. The method of claim 3, wherein determining answers to each of the high frequency questions based on the matching results comprises:
and according to the matching result, taking the intermediate information text matched with the standard answer in the intermediate information text of each high-frequency question as the answer of the corresponding high-frequency question, removing the intermediate information text which is not matched with the standard answer in the intermediate information text of each high-frequency question, and taking the standard answer corresponding to the high-frequency question with the removed intermediate information text as the answer.
5. The method of claim 1, wherein the obtaining chat detail information generated during interaction between the e-commerce customer service and the customer of each virtual store in the vendor workstation application comprises:
starting a seller workbench application window;
entering a chat window interface of each virtual store;
and acquiring webpage data on a chat window interface of each shop, and taking the acquired webpage data as chat detail information.
6. The method of claim 5, wherein the collecting the webpage data on the chat window interface of each store, and the using the collected webpage data as the chat detail information comprises:
acquiring all target links of a current webpage;
clicking all the target links, and entering a detail page corresponding to each target link;
extracting data in a detail page corresponding to each target link, and taking the extracted data in each detail page as chatting detail data;
and searching whether a page turning button exists in the current webpage or not, if so, clicking the page turning button to enter the next webpage, taking the next webpage as the current webpage, and returning to the step of acquiring all target links of the current webpage until the page turning button cannot be searched.
7. The method of claim 5, wherein the launching the vendor workstation application window comprises:
starting a browser driver using an automated testing tool to start a vendor workstation application;
and starting an application window of the seller workbench application through remote driving.
8. A text processing apparatus based on RPA and AI, comprising:
the information acquisition unit is used for acquiring chatting detail information of each shop in a seller workbench application, wherein the chatting detail information is generated in the interaction process of an e-commerce customer service and a customer, and the chatting detail information of each shop comprises a plurality of questions and original information texts corresponding to the questions;
the high-frequency question acquisition unit is used for performing text recognition and data statistics on each question in the chat detail information to obtain a plurality of high-frequency questions;
the answer obtaining unit is used for extracting information and performing text emotion analysis on the original information text of each high-frequency question to obtain an answer of the high-frequency question;
the text generation unit is used for combining each high-frequency question and the answer of the high-frequency question into a target text;
and the text writing unit is used for writing all the target texts into a background QA knowledge base.
9. A text processing apparatus based on RPA and AI, comprising: a memory and a processor;
the memory is used for storing programs;
the processor is configured to execute the program to implement the steps of the RPA and AI based text processing method according to any one of claims 1 to 7.
10. A storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the steps of the RPA and AI-based text processing method according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210750317.1A CN115080720A (en) | 2022-06-29 | 2022-06-29 | Text processing method, device, equipment and medium based on RPA and AI |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210750317.1A CN115080720A (en) | 2022-06-29 | 2022-06-29 | Text processing method, device, equipment and medium based on RPA and AI |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115080720A true CN115080720A (en) | 2022-09-20 |
Family
ID=83255373
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210750317.1A Pending CN115080720A (en) | 2022-06-29 | 2022-06-29 | Text processing method, device, equipment and medium based on RPA and AI |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115080720A (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105893344A (en) * | 2016-03-28 | 2016-08-24 | 北京京东尚科信息技术有限公司 | User semantic sentiment analysis-based response method and device |
CN111324704A (en) * | 2018-12-14 | 2020-06-23 | 阿里巴巴集团控股有限公司 | Method and device for constructing dialect knowledge base and customer service robot |
CN112148846A (en) * | 2020-08-25 | 2020-12-29 | 北京来也网络科技有限公司 | Reply voice determination method, device, equipment and storage medium combining RPA and AI |
CN112287082A (en) * | 2019-12-31 | 2021-01-29 | 北京来也网络科技有限公司 | Data processing method, device, equipment and storage medium combining RPA and AI |
CN112818102A (en) * | 2021-02-01 | 2021-05-18 | 杭州微洱网络科技有限公司 | Context-based fast question answering method for FAQ (failure of expert knowledge) knowledge base |
CN114186041A (en) * | 2021-11-29 | 2022-03-15 | 珠海采筑电子商务有限公司 | Answer output method |
CN114385612A (en) * | 2021-12-29 | 2022-04-22 | 深圳市信联征信有限公司 | Data processing method, data display method, data processing device, data display device, equipment and storage medium |
-
2022
- 2022-06-29 CN CN202210750317.1A patent/CN115080720A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105893344A (en) * | 2016-03-28 | 2016-08-24 | 北京京东尚科信息技术有限公司 | User semantic sentiment analysis-based response method and device |
CN111324704A (en) * | 2018-12-14 | 2020-06-23 | 阿里巴巴集团控股有限公司 | Method and device for constructing dialect knowledge base and customer service robot |
CN112287082A (en) * | 2019-12-31 | 2021-01-29 | 北京来也网络科技有限公司 | Data processing method, device, equipment and storage medium combining RPA and AI |
CN112148846A (en) * | 2020-08-25 | 2020-12-29 | 北京来也网络科技有限公司 | Reply voice determination method, device, equipment and storage medium combining RPA and AI |
CN112818102A (en) * | 2021-02-01 | 2021-05-18 | 杭州微洱网络科技有限公司 | Context-based fast question answering method for FAQ (failure of expert knowledge) knowledge base |
CN114186041A (en) * | 2021-11-29 | 2022-03-15 | 珠海采筑电子商务有限公司 | Answer output method |
CN114385612A (en) * | 2021-12-29 | 2022-04-22 | 深圳市信联征信有限公司 | Data processing method, data display method, data processing device, data display device, equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP7042693B2 (en) | Interactive business support system | |
US9104709B2 (en) | Cleansing a database system to improve data quality | |
JP2005222532A (en) | Machine learning approach for determining document relevance for searching large-scale collection of electronic document | |
EP3685245B1 (en) | Method, apparatus, and computer-readable media for customer interaction semantic annotation and analytics | |
US8560518B2 (en) | Method and apparatus for building sales tools by mining data from websites | |
CN108647729B (en) | User portrait acquisition method | |
CN108345663A (en) | A kind of news push method and apparatus | |
CA3153056A1 (en) | Intelligently questioning and answering method, device, computer, equipment and storage medium | |
JP2005190284A (en) | Information classification device and method | |
CN116452212B (en) | Intelligent customer service commodity knowledge base information management method and system | |
CN112749238A (en) | Search ranking method and device, electronic equipment and computer-readable storage medium | |
CN116703515A (en) | Recommendation method and device based on artificial intelligence, computer equipment and storage medium | |
CN115080720A (en) | Text processing method, device, equipment and medium based on RPA and AI | |
CN112148939A (en) | Data processing method and device and electronic equipment | |
CN112328812A (en) | Domain knowledge extraction method and system based on self-adjusting parameters and electronic equipment | |
CN112771508A (en) | Seminar auxiliary system and seminar auxiliary method | |
US11914844B2 (en) | Automated processing and dynamic filtering of content for display | |
CN116775813B (en) | Service searching method, device, electronic equipment and readable storage medium | |
Suleiman et al. | Process Mining Enabled Cognitive RPA to Automate Data Entry Tasks in ERP Systems | |
CN118210977A (en) | Object recommendation method and device based on session data, electronic equipment and medium | |
CN113326062A (en) | Software defect-oriented multi-round automatic question and answer method, system, computer equipment and storage medium | |
CN116703416A (en) | Method, system, apparatus and medium for processing user feedback | |
CN117667979A (en) | Data mining method, device, equipment and medium based on large language model | |
Sharda et al. | Information extraction from interviews to obtain tacit knowledge: A text mining application | |
CN117333191A (en) | Complaint event association method and device and complaint event association system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |