CN110807091B - Hotel intelligent question-answer recommendation and decision support analysis method and system - Google Patents

Hotel intelligent question-answer recommendation and decision support analysis method and system Download PDF

Info

Publication number
CN110807091B
CN110807091B CN201910154681.XA CN201910154681A CN110807091B CN 110807091 B CN110807091 B CN 110807091B CN 201910154681 A CN201910154681 A CN 201910154681A CN 110807091 B CN110807091 B CN 110807091B
Authority
CN
China
Prior art keywords
hotel
customer
database
entity
attribute
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.)
Active
Application number
CN201910154681.XA
Other languages
Chinese (zh)
Other versions
CN110807091A (en
Inventor
王涵
黄业坚
黄国琼
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Individual
Original Assignee
Individual
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Individual filed Critical Individual
Priority to CN201910154681.XA priority Critical patent/CN110807091B/en
Publication of CN110807091A publication Critical patent/CN110807091A/en
Application granted granted Critical
Publication of CN110807091B publication Critical patent/CN110807091B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0631—Item recommendations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/12—Hotels or restaurants
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

The invention relates to a hotel intelligent question-answer recommendation and decision support analysis method and a system, which are used for realizing: the crawler is utilized to quickly and efficiently crawl mass data of a plurality of tourist websites and update the mass data in real time; extracting comment information hotel attribute keywords by using an algorithm, and carrying out emotion analysis and partitioning on each keyword; arranging a related entity database, an attribute database and a client-hotel relation database, and constructing a knowledge graph; a question-answering template is written by using a knowledge graph tool and a query language, and a message is identified and automatically answered by using a natural language processing technology and an entity identification technology. The beneficial effects of the invention are as follows: learning customer preference, and intelligently recommending hotels to customers to automatically reply to customer comment information, and simultaneously realizing targeted transmission of hotel advertisement consultation in a mail or message form; the system is optimized continuously by oneself, question and answer templates and intelligent sequencing are enriched, and the system is more civilian and available.

Description

Hotel intelligent question-answer recommendation and decision support analysis method and system
Technical Field
The invention relates to a hotel intelligent question-answer recommendation and decision support analysis method and system, and belongs to the field of computers.
Background
With the increasing proportion of social media networks in customer decisions, various travel modes such as business travel, home travel, academic travel, medical travel, rural travel and the like go deep into the mind, and the ordering of travel subsidiary consumer products such as air tickets, civilian destinations, hotels and entrance tickets through travel websites has become the main stream of travel product sales. The online combined marketing mode makes travel more convenient and faster. The comment information of the tourist website plays a great role in the commodity purchasing process of the customer, and the comment information is effectively mined, so that the disordered comment information of the customer is organized into a searchable and structured database, the useful information of the social media network is organized into a structured knowledge network, the true and reliable purchasing recommendation is provided for the customer, the effective decision support analysis is provided for the merchant, and the intelligent interaction among the comment information knowledge base, the customer and the merchant is realized. For this reason, the advent of artificial intelligence technology has provided technical possibilities for this purpose for the present invention. The invention is based on diversity, reality, real-time and huge amount of travel website information, and based on knowledge graph, self-recognition language processing and emotion analysis artificial intelligence technology, achieves hotel intelligent reply, hotel intelligent recommendation and decision support analysis, thereby saving time for browsing comment information of each merchant for customers, providing real-time decision support for merchants, effectively promoting merchants to achieve customer relationship management, customer satisfaction and loyalty evaluation, effectively directing personalized promotion dynamic messages and the like.
The number of tourist websites in the world is not good, and with the rising of tourist heat in the country, hotels where the tourist websites stay go through the management ball and all places, so as to realize the comprehensive digging of available information of each website, and a relatively complete real-time knowledge graph is quickly built.
Disclosure of Invention
The invention provides a hotel intelligent question-answer recommendation and decision support analysis method and system. The invention firstly constructs a multi-thread and multi-process high-speed crawler to crawl hotel social media network customers, hotel attribute information and hotel and customer relation information, such as reviewer ID, check-in time, housing type, travel type, birth place, hotel star grade, affiliated group, comment information, scoring information and the like. Then, using a jeeba, google and other public Word banks to perform Word segmentation and stop Word removal, adopting a genesim, word2Vec, sklearn, numpy, collections, math and other related Python databases to generate Word vector matrixes, adopting and not only TextRank, LDA, TF-IDF, TPR and a random forest office decision tree (keywords: predictive category 1, non-keywords: predictive category 2) to extract keywords (predictive category 1), using the separator to generate phrases, and calculating emotion scores of the keywords by taking the phrases as units (the emotion scores are continuously distributed in [0,1], "1" is positive emotion score, and "0" is negative emotion score). The invention sorts the database into an entity database containing attributes and an entity relationship database. The invention uses entity data as nodes, relation data as node relation, builds knowledge graph based on but not limited to Neo4J knowledge graph building tool and Python programming language, uses py2Neo to build knowledge graph, and designs problem template for three aspects of entity, attribute and relation, for the system user to inquire the entity, attribute and relation. In addition, the invention designs a question input box, and uses natural language processing technology to identify condition sentences and inquiry sentences, and identify entity nodes, attributes and correlations of the condition sentences and the inquiry sentences, so as to realize the scene constructed by a user of the recognition system, understand the context and the content of the questions, and answer questions presented by the user according to the constructed knowledge graph artificial intelligence. Meanwhile, according to customers carrying different attributes, the first three hotels with the highest emotion average of the attributes of the concerned hotels are recommended based on the attributes of the hotels concerned by the customers. In addition, the system automatically identifies the comment content of the customer by using a natural language processing technology, edits and sets automatic replies, carries out natural language processing on the advertisement consultation content issued by the hotel, identifies the entity and the related attribute thereof, and automatically sends and recommends different customers in the forms of mail, message automatic replies and the like.
The technical scheme of the invention comprises a hotel intelligent question-answer recommendation and decision support analysis method, which is characterized by comprising the following steps: s100, compiling and using a crawler to crawl relationship data of a hotel and a customer from a hotel social media network; s200, extracting hotel attribute keywords from the comment information, and carrying out emotion analysis and scoring on each attribute keyword; s300, creating and arranging an entity database, a hotel attribute database, a customer attribute database and an entity relation database, and outputting a result; s400, integrating the results output by the S300, taking the integrated results as knowledge graph input data, and establishing a knowledge graph of a customer and a hotel; s500, creating an intelligent question-answer template for the frequently-used questions and answers of the hotel and the customer according to the database obtained in the S300; s600, creating a recommendation module for hotel and customer attributes according to the database obtained in the S300; s700, compiling hotel and customer attributes into a decision support analysis module according to the database obtained in the S300; s800, optimizing the case record of the system usage generated in S500 using machine learning.
According to the hotel intelligent question-answer recommendation and decision support analysis method, the S100 specifically comprises the following steps: writing and using a high-speed crawler to crawl the relationship data of customers, hotels and the customers and the hotels from the hotel social media network; wherein the customer data includes, but is not limited to, comment ID, time of stay, housing type, travel type, place of birth; information of hotels includes, but is not limited to, hotel stars and affiliated groups; hotel and customer relationship data includes, but is not limited to, comment information and scoring information; the crawled information includes structural, semi-structural data, and non-structural data.
According to the hotel intelligent question-answer recommendation and decision support analysis method, the S200 specifically comprises the following steps: s201, importing a hotel special word database, and using Guan Yuliao libraries such as Jieba, google English corpus and the like to cut words from the crawled comment data and remove stop words; s202, using algorithms such as TextRank, LDA, TF-IDF, TPR and the like, generating Word vector matrixes by using a gensim, word2Vec, sklearn, numpy, collections and a math Python database, extracting keywords by adopting a decision tree not limited to a random forest bureau, calculating accuracy and recall rate, calculating an average value, and arranging and outputting; the accuracy rate refers to the percentage of the predicted 1 keyword to the predicted keyword, the recall rate refers to the percentage of the predicted 1 keyword to the original keyword, the predicted 1 represents the keyword prediction category, and the predicted 2 represents the non-keyword prediction category; s203, locating around each keyword by using word vectors, dividing the keyword into a plurality of short sentences by using a separator, identifying and counting emotion words around the keywords by using the short sentences as units, marking positive emotion words as 1, marking negative emotion words as 0, obtaining emotion average of the whole comment on different keywords, finally obtaining average emotion score of each customer on the hotel, using the emotion score as a percentage as a recommendation index of the customer on the hotel, dispersing the score in a range of [0,0.5 ] and [0.5,1], and marking that the relation result does not recommend 0 and recommendation 1.
According to the hotel intelligent question-answer recommendation and decision support analysis method, the S300 specifically comprises the following steps: s301, arranging an entity database containing attributes, wherein the entity database comprises: the hotel attribute database comprises attributes such as hotel stars, affiliated groups and the like, and the customer attribute database comprises keywords such as customer comment IDs, ages, sexes, birth places, travel types, houses and focused hotel attributes; s302, arranging an entity relation database, wherein the entity relation database comprises: the relationship among the customers, the relationship among the hotels and the relationship among the customers and the hotels, wherein the relationship data among the customers comprises, but is not limited to, the same age, the same sex, the same house, the same country and the same mention of certain hotel attribute keywords, and the hotel relationship data comprises, but is not limited to, the same star class and the same group flag; the customer-to-hotel relationship includes whether the customer recommends a hotel or not, and a recommendation index.
According to the hotel intelligent question-answer recommendation and decision support analysis method, the S400 specifically comprises the following steps: entity data is used as nodes and relationship data is used as node relationships, including but not limited to knowledge graph building using Neo4J knowledge graph building tool, python programming language, and py2 Neo.
According to the hotel intelligent question-answer recommendation and decision support analysis method, S500 specifically comprises: s501, setting a question template according to the database obtained in S300, using py2neo and other related Python libraries, establishing attribute, entity and relation synonym libraries, writing common questions into SPARQL, cypher query sentences, establishing common question-answering templates, selecting question-answering forms, answering the common questions, including but not limited to hotel customer number columns conforming to the attribute or relation, inquiring the relation among the entities, and the like, and intelligently sequencing the templates through a machine learning system user query record; s502, providing a question input box for inputting a message, wherein the message comprises but is not limited to hotel experience, expected expectations of customers on hotels and target object inquiry of hotel managers for hotel special information pushing, identifying the input question by using natural language processing, identifying a conditional sentence and an inquiry sentence, respectively identifying entity nodes, attributes and correlations of the conditional sentence and the inquiry sentence, and outputting inquiry contents according to the inquiry sentence of S501; s503, based on the neural network deep learning related models such as LSTM, RNN and the like, using but not limited to CRF++, neuroNER and the like, carrying out named entity identification, labeling entity nodes, attributes and interrelationships, converting a problem into a mode of 'condition input-result inquiry', writing the common problem of the mode into a SPARQL, cypher inquiry statement, establishing questions and answers and outputting results, and simultaneously adding the mode of 'condition input-result inquiry' of message information into a system user record, carrying out machine learning, optimizing templates and intelligently sequencing.
According to the hotel intelligent question-answer recommendation and decision support analysis method, the S600 specifically comprises the following steps: s601, recommending the first three hotels with the highest average emotion score of the hotel attributes in a specific area based on the hotel attributes focused by the customers according to the customers with different attributes; s602, automatically identifying customer comment content and editing and setting automatic reply by using a natural language processing technology, performing natural language processing on advertisement consultation content issued by a hotel, identifying an entity and related attributes thereof, and automatically sending and recommending different customers in the forms of mail, message automatic reply and the like.
According to the hotel intelligent question-answer recommendation and decision support analysis method, the S700 specifically comprises the following steps: word cloud display is carried out on keywords, descriptive analysis, information mining and graphic visualization are carried out on emotion analysis results, the emotion scores of all hotel attributes are predicted by using but not limited to LSTM and ARMA algorithms, relevance exploration is carried out by using but not limited to Apriori, xgboost, and decision support data are provided for different hotels.
According to the hotel intelligent question-answer recommendation and decision support analysis method, the S800 specifically comprises: s801, machine learning is carried out on the system usage records generated in S500, templates are continuously enriched and reordered, so that the system is more intelligent and easier to operate, namely, machine learning is carried out on the system usage record database generated in S500, and intelligent ordering of templates is optimized.
The technical scheme of the invention also comprises a hotel intelligent question-answer recommendation and decision support analysis system which is used for executing any method, and the system comprises the following steps: the data acquisition module is used for compiling and using a crawler to crawl the relationship data of the hotel and the customer from the hotel social media network; the keyword extraction and emotion analysis module is used for extracting hotel attribute keywords from the comment information, and carrying out emotion analysis and scoring on each attribute keyword; the knowledge graph database preparation module is used for creating and arranging an entity database, a hotel attribute database, a customer attribute database and an entity relation database, and outputting a result; the knowledge graph establishing module is used for integrating the results output by the knowledge graph database preparing module, taking the integrated results as knowledge graph input data and establishing the knowledge graphs of customers and hotels; the intelligent question-answering module is used for intelligently answering common questions and complex questions of the clients and hotel managers; the intelligent recommending module is used for recommending relevant hotels and realizing the pushing and automatic replying of the hotels to the personalized advertisement consultation of the customers; the decision support analysis module is used for providing decision support and data analysis; the system optimization module is used for optimizing the hotel intelligent question-answer recommendation and decision support analysis system based on knowledge graph, natural language processing and emotion analysis.
The beneficial effects of the invention are as follows: the multi-thread and multi-process crawler is utilized to quickly and efficiently crawl mass data of a plurality of tourist websites and update the mass data in real time; extracting comment information hotel attribute keywords by using an algorithm, carrying out emotion analysis on each keyword, and distributing scoring on the intervals of [0,1 ]; arranging a related entity database, an attribute database and a client-hotel relation database, and constructing a client-hotel knowledge graph; writing a question-answering template by using a knowledge graph tool and a query language, and identifying a message and automatically answering a question of a system user by using a natural language processing technology and an entity identification technology; through artificial intelligence technology, learning customer preference, recommending hotel for customer individuation, and using natural language processing technology and entity recognition technology for customer comment information and hotel advertisement consultation information, automatically replying only to customer comment information, and simultaneously realizing directional transmission of hotel advertisement consultation in mail or message form; meanwhile, the system continuously optimizes the system by machine learning processing of the system usage record database, enriches question-answering templates and intelligently sorts, so that the system is more civilian and available.
Drawings
FIG. 1 is a general flow chart of a method according to an embodiment of the invention;
FIG. 2 is a block diagram of a system architecture according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a "multithreading+multiprocessing" travel website high-speed crawler according to an embodiment of the invention;
FIG. 4 is a schematic diagram of a one-stop crawler flow according to an embodiment of the present invention;
fig. 5 is a schematic flow chart of a method of the intelligent question-answering module according to an embodiment of the invention.
Detailed Description
The technical scheme of the invention provides a hotel intelligent question-answer recommending and decision support analyzing method and system, which relate to technologies such as web crawlers, parallel computing, machine learning, knowledge graphs, database management, recommending models, predicting models, data mining and the like, and the conception, specific structure and technical effects generated by the method are clearly and completely described by combining the embodiments and the drawings so as to fully understand the purposes, the schemes and the effects of the method.
It should be noted that, unless otherwise specified, when a feature is referred to as being "fixed" or "connected" to another feature, it may be directly or indirectly fixed or connected to the other feature. Further, the descriptions of the upper, lower, left, right, etc. used in this disclosure are merely with respect to the mutual positional relationship of the various components of this disclosure in the drawings. As used in this disclosure, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. Furthermore, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in the description presented herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and/or" as used herein includes any combination of one or more of the associated listed items.
It should be understood that although the terms first, second, third, etc. may be used in this disclosure to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element of the same type from another. For example, a first element could also be termed a second element, and, similarly, a second element could also be termed a first element, without departing from the scope of the present disclosure. The use of any and all examples, or exemplary language (e.g., "such as") provided herein, is intended merely to better illuminate embodiments of the invention and does not pose a limitation on the scope of the invention unless otherwise claimed.
Fig. 1 shows a general flow chart of a method according to an embodiment of the invention. The method comprises the following steps:
(1) The travel website collects data: the related program packages such as Python programming language, datetime, pymongo, pysql, download, mongodb _ queue, threading, multiprocessing, bs, re, selenium and the like are used for designing and constructing a multi-thread and multi-process universal high-speed crawler for a travel website, and related data of customers, hotels and the relationship between the customers and the hotels are crawled from a plurality of platforms of a hotel social media network. Customer data includes, but is not limited to, comment ID, time of stay, housing type, travel type, place of birth, etc.; information of hotels includes, but is not limited to, hotel stars, affiliated groups, etc.; hotel and customer relationship data includes, but is not limited to, comment information, scoring information, and the like. The crawling information comprises structural data, semi-structural data and non-structural data. For example: the method comprises the steps of crawling Chinese and English data of 20 hotels in 2017 in the first rank of a certain area of three travel websites such as the travel website 1, the travel website 2 and the travel website 3, running processes on 3 different cores of CPU respectively, storing websites in the process in a website queue to be crawled, a website queue being crawled, a website queue crawled, and crawling information such as reviewer ID, reviewer birth place, reviewer age, reviewer room time, comment title, comment content, hotel star level, hotel affiliated group and the like in each process.
Fig. 3 is a schematic diagram of a high-speed crawler of a "multithreading+multiprocessing" travel website, and fig. 4 is a schematic diagram of a one-stop crawler flow: sorting to-be-crawled hotel lists, inputting hotel names into different travel websites, constructing url lists, taking each travel website as a process, sorting the to-be-crawled website queues, the forward crawled website queues and the crawled website queues in a queue sorting mode in each process, setting reasonable Timeout as a threshold value, placing websites with processing time exceeding the set Timeout in the to-be-crawled website queues, and placing successfully crawled website addresses in the crawled website queues. And the process of crawling target data is used as a daemon thread, and the crawled data is stored in a database. In this regard, different processes may run on different cores of the cup, and different computers may be implemented online or cloud crawling. The one-stop model of "simulating browser-obtaining web page code-crawling required data-simulating clicking next page-closing browser and storing crawling data" can be used for the models adopted in the crawling process of different threads.
(2) Keyword extraction and emotion analysis: importing a hotel special word database, and using Guan Yuliao libraries such as Jieba, google English corpus and the like to cut words from the crawled comment data and remove stop words; the method comprises the steps of generating a Word vector matrix by using algorithms such as TextRank, LDA, TF-IDF, TPR and the like and related Python databases such as gensim, word2Vec, sklearn, numpy, collections, math and the like, extracting keywords (prediction category 1) by adopting a decision tree (keywords: prediction category 1 and non-keywords: prediction category 2) of a random forest office, calculating accuracy and recall, calculating an average value and outputting the result in sequence. The accuracy rate refers to the percentage of the predicted 1 keyword to the predicted keyword, and the recall rate refers to the percentage of the predicted 1 keyword to the original keyword. The method comprises the steps of locating around each keyword by using word vectors, dividing the keyword into a plurality of short sentences by using a separator, identifying emotion words around the statistical keywords by taking the short sentences as a unit, marking positive emotion words as 1, marking negative emotion words as 0, averaging emotion of the whole comment on different keywords, finally obtaining average emotion score of each customer on the hotel, marking the emotion score as a percentage as a recommended index of the customer on the hotel, dispersing the score on a section of [0,0.5 ] and [0.5,1] and marking the score as a relation result [ not recommended 0, recommended 1]. For example, "banquet hall" and "dining" are two hotel attribute keywords, and if "a banquet hall decoration in a certain area is very bright and dining is possible, the banquet hall is not served", then the attention points of the customer are "banquet hall" and "dining", and the emotion scores of the two are "(1+0)/2=0.5" and "1", respectively.
(3) Sorting and establishing entity, attribute and relation databases: sorting an entity database containing attributes: the entity database comprises a hotel attribute database and a customer attribute database, wherein the hotel attribute database comprises attributes such as hotel stars, affiliated groups and the like, the customer attribute database comprises attributes such as customer comment IDs, ages, sexes, places of birth, travel types, housing types and focused hotel attribute keywords (attribute keywords are noted as 1 in comments and attribute keywords are not noted as 0); sorting an entity relationship database: the entity relation database comprises a relation among customers, a relation among hotels and a relation among customers and hotels, wherein the relation data among customers comprises, but is not limited to, the same age, the same sex, the same house type, the same country, the same mention of certain hotel attribute keywords (the relation is specific to a single age group, the sex, the house type, the place of birth, the attribute keywords and the like), and the hotel relation data comprises, but is not limited to, the same star level, the same group flag (the relation is specific to a certain star level, the group name and the like); the customer-to-hotel relationship includes whether the customer recommends a hotel or not, and a recommendation index. For example: the customer attribute database is: a customer: ID: mini, place of birth: chinese, room type: large bed room, injection point: "service"; customer B ID: ivanka, place of birth: chinese, room type: haze-free house and point injection: "services" and "rooms". Hotel attribute database: some hotels in some areas: star grade: 5, group: intercontinental, "service" emotion average score: 0.8, "room" emotion average: 0.9; somewhere in a certain area, somewhere in a certain hotel: star level: 5, group: some "service" emotion average score: 0.7, average "room" emotion score: 0.6. relational database: both the A and B customers are born in China, and pay attention to service; and a certain hotel in a certain area and a certain hotel in a certain place in a certain area belong to a certain group and are 5-star hotels.
(4) Building a 'customer-hotel' knowledge graph, and building a customer hotel common problem template and performing natural language processing intelligent question-answer search: using entity data as nodes, using relation data as node relations, and establishing a knowledge graph by using py2Neo based on a Neo4J knowledge graph construction tool and a Python programming language; the problem template may be: 1. those hotels that "serve" best in a certain area? 2. Where are customers focusing on "price" all come from? 3. ____ (where the number (1) service is filled, (2) room, (3) banquet hall, (4) facilities, etc., which of a plurality of worth recommending hotels is filled? 4. How many percent of customers are in ____ (where the number (1) service is filled, (2) room, (3) banquet hall, (4) facilities, etc. can be filled in a plurality)? Etc
FIG. 4 is a schematic diagram of a one-stop crawler flow according to an embodiment of the present invention.
Fig. 5 is a flow chart of a method of an intelligent question-answering module according to an embodiment of the present invention: s501, setting a question template according to the database obtained in S300, using py2neo and other related Python libraries, establishing attribute, entity and relation synonym libraries, writing common questions into SPARQL, cypher query sentences, establishing common question-answering templates, and selecting question-answering forms to answer the common questions, wherein the questions comprise but are not limited to hotel customer number columns conforming to the attribute or relation, relation query among entities and the like, and meanwhile, the system intelligently sorts the templates through a machine learning system user query record; providing a question input box, inputting a longer message, wherein the message contains rich information, which can be the previous hotel experience, the customer expects the message, a hotel manager uses natural language processing to identify the input question aiming at the inquiry of a target object of the hotel special message, and identifies a conditional sentence and an inquiry sentence, respectively identifies the entity node, attribute and interrelation of the conditional sentence and the inquiry sentence, and outputs inquiry content in combination with an S501 inquiry sentence; based on the neural network deep learning related models such as LSTM, RNN and the like, using but not limited to CRF++, neuroNER and the like, carrying out named entity identification, labeling entity nodes, attributes and interrelationships, converting a problem into a mode of 'condition input-result inquiry', writing the common problem of the mode into a SPARQL, cypher inquiry statement, establishing questions and answers and outputting results, and simultaneously adding a mode of 'condition input-result inquiry' of message information into a system user record, carrying out machine learning, optimizing a template and intelligently sequencing. The problem template may be: 1. those hotels that "serve" best in a certain area? 2. Where are customers focusing on "price" all come from? 3. ____ (where the number (1) service is filled, (2) room, (3) banquet hall, (4) facilities, etc., which of a plurality of worth recommending hotels is filled? 4. How many percent of customers are in ____ (where the number (1) service is filled, (2) room, (3) banquet hall, (4) facilities, etc. can be filled in a plurality)? Etc.
(5) According to the hotel attribute emotion score and the customer attention attribute, realizing individual recommendation, hotel automatic reply and advertisement consultation pushing: recommending the first three hotels with highest average emotion scores of the hotel attributes in the specific area based on the hotel attributes focused by the customers according to the customers with different attributes; the method comprises the steps of automatically identifying customer comment content and editing and setting automatic reply by using a natural language processing technology, carrying out natural language processing on advertisement consultation content issued by a hotel, identifying entities and related attributes thereof, and automatically sending and recommending different customers in the forms of mail, message automatic reply and the like. For example: some comments "today, when I come out of the bathroom, the floor is very slippery, and I beat a foot, and are good for mind. "entity word: "bathroom", "floor", "bathroom lacks emotion word", "floor" emotion score is 0, the total emotion score is 0, then the system replies automatically: "very sorry you are dissatisfied with our floors, we will deal with immediately, hope you to have a pleasant recall next time"; the latest marketing consultation of a hotel in a place is "during the spring festival, a western restaurant sells a 99 generation 200 generation gold coupon", the entity word of the consultation is "western restaurant" and "price", and then the message is directed to a customer focusing on the "western restaurant" and "price".
(6) Performing time sequence prediction and correlation analysis on the attribute emotion analysis score, and providing decision support for a decision maker: word cloud display is carried out on keywords, descriptive analysis, information mining and graphic visualization are carried out on emotion analysis results, the emotion scores of all hotel attributes are predicted by using but not limited to LSTM and ARMA algorithms, relevance exploration is carried out by using but not limited to Apriori, xgboost, and decision support data is provided for different hotels;
(7) Machine learning system usage record, implementation system optimization: and (3) machine learning is carried out on the system usage records generated in the step S500, templates are continuously enriched and reordered, so that the system is more intelligent and easier to operate, namely, machine learning is carried out on the system usage record database generated in the step S500, and intelligent ordering of the templates is optimized.
The technical scheme of the invention specifically discloses a more detailed implementation scheme, which is specifically as follows:
(1) The example uses Python programming language, datetime, pymongo, pysql, download, mongoDB _ queue, threading, multiprocessing, bs, re, selenium and other relevant program packages to design and write a multi-thread and multi-process travel website high-speed crawler to crawl comments and customer information of a hotel 2017 month 8 to 2018 month 8 travel website 1, travel website 2 and other certain websites in a certain area, uses a Pycham IDE compiler, mongoDB as a basic data storage unit, adopts 4-core processor computer Ethernet as a crawling environment, and crawls 49702 pieces of comment information altogether, and takes 3 hours and 42 minutes.
(2) According to the invention, a Jieba and a certain English corpus are used, python is adopted for writing, related Python databases such as gensim, word2Vec, sklearn, numpy, collections, math and the like are used, after stop words are removed, a TF-ITF algorithm is adopted, the first 600 entity keywords are extracted, 25 hotel attribute keywords are obtained through synonym replacement, and emotion scoring is carried out on 25 attributes of the 28 hotels by using an emotion analysis model.
(3) The databases are organized into a customer-attribute database, a hotel-attribute database and a relational database, wherein 43055 customers do not repeatedly comment on ID, 4 customers have the attributes of 'birth place', 'house', 'play mode', 'comment date/check-in date', 29 hotel attributes have the average emotion score of 25 attributes, average emotion score of total emotion score, star grade, region and group.
(4) According to the invention, 43055 'customers' and 28 'hotels' are taken as entity nodes, customer attributes and hotel attributes are taken as related attribute characteristics, a Neo4J knowledge graph construction tool and a Python programming language are used, and py2Neo is imported to establish a knowledge graph model.
(5) Setting up a problem template: 1. those hotels that "serve" best in a certain area? 2. Where are customers focusing on "price" all come from? 3. ____ (where the number (1) service is filled, (2) room, (3) banquet hall, (4) facilities, etc., which of a plurality of worth recommending hotels is filled? 4. How many percent of customers are in ____ (where the number (1) service is filled, (2) room, (3) banquet hall, (4) facilities, etc. can be filled in a plurality)? Using py2neo and other related Python libraries to build attribute, entity and relationship synonym libraries, writing common questions into SPARQL query sentences, building common question-answering templates, selecting question-answering forms, answering common questions, realizing hotel customer series of attribute or relationship, querying relationship among entities and the like.
(6) Based on the RNN neural network deep learning related model, using the neuroNER to conduct named entity identification, labeling entity nodes, attributes and interrelationships, converting a problem into a mode of 'condition input-result query', writing the common problem in the mode into a SPARQL query statement, establishing a question and answer, and outputting a result.
(7) Planning (5) the system usage record as a pattern of "conditional input-result query", for example: a system user inputs in an input box: "some flag service in some area is better than a hotel, where general customers are more", the system converts it into: conditions are as follows: hotel address: somewhere, "service" emotion analysis score > =0.5, hotel group: some; inquiring: is the customer's place of birth? By "put it into the system usage record database and add it to the template, cluster and reorder the system usage record data, put the class selected to contain the query at the top of the user question template. Thereby achieving the optimization of the whole system.
The present invention is not limited to the above embodiments, but can be modified, equivalent, improved, etc. by the same means to achieve the technical effects of the present invention, which are included in the spirit and principle of the present invention. Various modifications and variations are possible in the technical solution and/or in the embodiments within the scope of the invention.

Claims (7)

1. The intelligent hotel question-answer recommending and decision support analyzing method is characterized by comprising the following steps of:
s100, crawling relationship data of a hotel and a customer from a hotel social media network by using a crawler;
s200, extracting hotel attribute keywords from the comment information, and carrying out emotion analysis and scoring on each attribute keyword;
s300, creating and arranging an entity database, a hotel attribute database, a customer attribute database and an entity relation database, and outputting a result;
s400, integrating the results output by the S300, taking the integrated results as knowledge graph input data, and establishing a knowledge graph of a customer and a hotel;
s500, creating an intelligent question-answer template for the frequently-used questions and answers of the hotel and the customer according to the database obtained in the S300;
s600, creating a recommendation module for hotel and customer attributes according to the database obtained in the S300;
s700, compiling hotel and customer attributes into a decision support analysis module according to the database obtained in the S300;
s800, optimizing the condition record of the system use generated in S500 by using machine learning;
wherein, the S300 specifically includes:
s301, arranging an entity database containing attributes, wherein the entity database comprises:
hotel attribute databases and customer attribute databases, wherein the hotel attribute databases include but are not limited to attributes of hotel stars, affiliated groups, etc., the customer attribute databases include but are not limited to customer comment IDs, ages, sexes, birth places, travel types, resident house type, focused hotel attribute keywords,
s302, arranging an entity relation database, wherein the entity relation database comprises:
the relationship between customers, the relationship between hotels and the relationship between customers and hotels, wherein the relationship data between customers comprises, but is not limited to, the same age, the same sex, the same house, the same country, the same mention of certain hotel attribute keywords, the hotel relationship data comprises, but is not limited to, the same star class, the same group flag,
the relationship between the customer and the hotel comprises whether the customer recommends the hotel or not and a recommendation index;
the step S500 specifically includes:
s501, setting a question template according to the database obtained in S300, using py2neo and other related Python libraries, establishing attribute, entity and relation synonym libraries, writing common questions into SPARQL, cypher query sentences, establishing common question-answering templates, having selected question-answering forms, answering common questions including but not limited to hotel customer number columns conforming to attributes or relations, entity-to-entity relation queries and the like, intelligently ordering the templates through a machine learning system user query record,
s502, providing a question input box for inputting a message, wherein the message comprises but is not limited to hotel experience, expected expectations of customers on hotels and target object inquiry of hotel managers for hotel special news pushing, identifying the input question by using natural language processing, identifying a conditional sentence and an inquiry sentence, respectively identifying entity nodes, attributes and correlations of the conditional sentence and the inquiry sentence, outputting inquiry contents according to the inquiry sentence of S501,
s503, based on the neural network deep learning related models such as LSTM, RNN and the like, using but not limited to CRF++, neuroNER to perform named entity recognition, labeling entity nodes, attributes and interrelationships, converting the questions into a mode of 'condition input-result inquiry', writing the common questions of the mode into SPARQL, cypher inquiry sentences, establishing questions and answers and outputting results, and simultaneously adding a mode of 'condition input-result inquiry' of message information into a system user record, performing machine learning, optimizing templates and intelligently sequencing;
the S600 specifically includes:
s601, recommending the first three hotels with the highest average emotion score of the hotel attributes in the specific area based on the hotel attributes focused by the customers according to the customers with different attributes,
s602, automatically identifying customer comment content and editing and setting automatic reply by using a natural language processing technology, performing natural language processing on advertisement consultation content issued by a hotel, identifying an entity and related attributes thereof, and automatically sending and recommending different customers in the forms of mail, message automatic reply and the like.
2. The hotel intelligent question-answer recommendation and decision support analysis method according to claim 1, wherein the S100 specifically comprises:
writing and using a high-speed crawler to crawl the relationship data of customers, hotels and the customers and the hotels from the hotel social media network;
wherein the customer data includes, but is not limited to, comment ID, time of stay, housing type, travel type, place of birth;
information of hotels includes, but is not limited to, hotel stars and affiliated groups;
hotel and customer relationship data includes, but is not limited to, comment information and scoring information;
the crawled information includes structural, semi-structural data, and non-structural data.
3. The hotel intelligent question-answer recommendation and decision support analysis method according to claim 1, wherein S200 specifically comprises:
s201, importing a hotel special word database, and using Guan Yuliao libraries such as Jieba, google English corpus and the like to cut words from the crawled comment data and remove stop words;
s202, using algorithms such as TextRank, LDA, TF-IDF, TPR and the like, generating Word vector matrixes by using a gensim, word2Vec, sklearn, numpy, collections and a math Python database, extracting keywords by adopting a decision tree not limited to a random forest bureau, calculating accuracy and recall rate, calculating an average value, and arranging and outputting;
the accuracy rate refers to the percentage of the predicted 1 keyword to the predicted keyword, the recall rate refers to the percentage of the predicted 1 keyword to the original keyword, the predicted 1 represents the keyword prediction category, and the predicted 2 represents the non-keyword prediction category;
s203, locating around each keyword by using word vectors, dividing the keyword into a plurality of short sentences by using a separator, identifying and counting emotion words around the keywords by using the short sentences as units, marking positive emotion words as 1, marking negative emotion words as 0, obtaining emotion average of the whole comment on different keywords, finally obtaining average emotion score of each customer on the hotel, using the emotion score as a percentage as a recommendation index of the customer on the hotel, dispersing the score in a range of [0,0.5 ] and [0.5,1], and marking that the relation result does not recommend 0 and recommendation 1.
4. The hotel intelligent question-answer recommendation and decision support analysis method according to claim 1, wherein S400 specifically comprises:
entity data is used as nodes and relationship data is used as node relationships, including but not limited to knowledge graph building using Neo4J knowledge graph building tool, python programming language, and py2 Neo.
5. The hotel intelligent question-answer recommendation and decision support analysis method according to claim 1, wherein S700 specifically comprises:
word cloud display is carried out on keywords, descriptive analysis, information mining and graphic visualization are carried out on emotion analysis results, the emotion scores of all hotel attributes are predicted by using but not limited to LSTM and ARMA algorithms, relevance exploration is carried out by using but not limited to Apriori, xgboost, and decision support data are provided for different hotels.
6. The hotel intelligent question-answer recommendation and decision support analysis method according to claim 1, wherein the S800 specifically comprises:
s801, machine learning is carried out on the system usage records generated in S500, templates are continuously enriched and reordered, so that the system is more intelligent and easier to operate, namely, machine learning is carried out on the system usage record database generated in S500, and intelligent ordering of templates is optimized.
7. A hotel intelligent question-answer recommendation and decision support analysis system for performing any of the methods of claims 1-6, comprising:
the data acquisition module is used for compiling and using a crawler to crawl the relationship data of the hotel and the customer from the hotel social media network;
the key word extraction and emotion analysis module is used for extracting hotel attribute keywords from the comment information, and carrying out emotion analysis and scoring on each attribute keyword;
the knowledge graph database preparation module is used for creating and arranging an entity database, a hotel attribute database, a customer attribute database and an entity relation database, and outputting a result;
the knowledge graph establishing module is used for integrating the results output by the knowledge graph database preparing module, taking the integrated results as knowledge graph input data and establishing the knowledge graphs of customers and hotels;
the intelligent question-answering module is used for intelligently answering common questions and complex questions of the clients and hotel managers;
the intelligent recommending module is used for recommending relevant hotels and realizing the pushing and automatic replying of the hotels to the personalized advertisement consultation of the customers;
the decision support analysis module is used for providing decision support and data analysis;
the system optimization module is used for optimizing the hotel intelligent question-answer recommendation and decision support analysis system based on knowledge graph, natural language processing and emotion analysis.
CN201910154681.XA 2019-03-01 2019-03-01 Hotel intelligent question-answer recommendation and decision support analysis method and system Active CN110807091B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910154681.XA CN110807091B (en) 2019-03-01 2019-03-01 Hotel intelligent question-answer recommendation and decision support analysis method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910154681.XA CN110807091B (en) 2019-03-01 2019-03-01 Hotel intelligent question-answer recommendation and decision support analysis method and system

Publications (2)

Publication Number Publication Date
CN110807091A CN110807091A (en) 2020-02-18
CN110807091B true CN110807091B (en) 2023-08-18

Family

ID=69487332

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910154681.XA Active CN110807091B (en) 2019-03-01 2019-03-01 Hotel intelligent question-answer recommendation and decision support analysis method and system

Country Status (1)

Country Link
CN (1) CN110807091B (en)

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111339269B (en) * 2020-02-20 2023-09-26 来康科技有限责任公司 Knowledge graph question-answering training and application service system capable of automatically generating templates
CN111341456B (en) * 2020-02-21 2024-02-23 中南大学湘雅医院 Method and device for generating diabetic foot knowledge graph and readable storage medium
CN111831880A (en) * 2020-02-21 2020-10-27 桂林电子科技大学 Intelligent question and answer method based on micro hotel platform
CN111340385A (en) * 2020-03-10 2020-06-26 深圳华侨城创新研究院有限公司 Scientific measuring method for measuring joy index of tourist attraction
US11222358B2 (en) 2020-05-29 2022-01-11 At&T Intellectual Property I, L.P. Facilitation of artificial intelligence predictions of telecommunications customers
CN112199484A (en) * 2020-10-12 2021-01-08 上海伊巢网络科技有限公司 Artificial intelligence customer service reply system for hotel residents
CN112256852A (en) * 2020-10-28 2021-01-22 北京软通智慧城市科技有限公司 Scenic spot comment data processing method and device, electronic equipment and storage medium
CN113190664B (en) * 2021-04-25 2022-06-10 内蒙古工业大学 Dynamic decision-making method for replying emotion in multi-turn conversation
CN113158082B (en) * 2021-05-13 2023-01-17 和鸿广科技(上海)有限公司 Artificial intelligence-based media content reality degree analysis method
CN113641798B (en) * 2021-10-12 2022-02-08 成都晓多科技有限公司 Identification method and system for disruptive comments of merchants
CN116823321B (en) * 2023-07-06 2024-02-06 青岛酒店管理职业技术学院 Method and system for analyzing economic management data of electric business
CN117522533A (en) * 2024-01-08 2024-02-06 江西求是高等研究院 Hotel intelligent searching method and system

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102117467A (en) * 2011-03-03 2011-07-06 南通大学 Question and answer mode online shopping guide method
CN103246677A (en) * 2012-02-13 2013-08-14 广州淘信互联网科技有限公司 Search method and search system on basis of social intercourse
WO2017010652A1 (en) * 2015-07-15 2017-01-19 포항공과대학교 산학협력단 Automatic question and answer method and device therefor
CN107633084A (en) * 2017-09-28 2018-01-26 武汉虹旭信息技术有限责任公司 Based on the public sentiment managing and control system and its method from media
CN107958091A (en) * 2017-12-28 2018-04-24 北京贝塔智投科技有限公司 A kind of NLP artificial intelligence approaches and interactive system based on financial vertical knowledge mapping
WO2018072563A1 (en) * 2016-10-18 2018-04-26 中兴通讯股份有限公司 Knowledge graph creation method, device, and system
CN107992543A (en) * 2017-11-27 2018-05-04 上海智臻智能网络科技股份有限公司 Question and answer exchange method and device, computer equipment and computer-readable recording medium
CN108182262A (en) * 2018-01-04 2018-06-19 华侨大学 Intelligent Answer System construction method and system based on deep learning and knowledge mapping
CN109145102A (en) * 2018-09-06 2019-01-04 杭州安恒信息技术股份有限公司 Intelligent answer method and its knowledge mapping system constituting method, device, equipment
CN109388748A (en) * 2018-09-26 2019-02-26 深圳壹账通智能科技有限公司 A kind of answering method of comment information, storage medium and server

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014056093A1 (en) * 2012-10-11 2014-04-17 Orboros, Inc. Method and system for visualizing complex data via a multi-agent query engine
US10956948B2 (en) * 2015-11-09 2021-03-23 Anupam Madiratta System and method for hotel discovery and generating generalized reviews
US10394956B2 (en) * 2015-12-31 2019-08-27 Shanghai Xiaoi Robot Technology Co., Ltd. Methods, devices, and systems for constructing intelligent knowledge base

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102117467A (en) * 2011-03-03 2011-07-06 南通大学 Question and answer mode online shopping guide method
CN103246677A (en) * 2012-02-13 2013-08-14 广州淘信互联网科技有限公司 Search method and search system on basis of social intercourse
WO2017010652A1 (en) * 2015-07-15 2017-01-19 포항공과대학교 산학협력단 Automatic question and answer method and device therefor
WO2018072563A1 (en) * 2016-10-18 2018-04-26 中兴通讯股份有限公司 Knowledge graph creation method, device, and system
CN107633084A (en) * 2017-09-28 2018-01-26 武汉虹旭信息技术有限责任公司 Based on the public sentiment managing and control system and its method from media
CN107992543A (en) * 2017-11-27 2018-05-04 上海智臻智能网络科技股份有限公司 Question and answer exchange method and device, computer equipment and computer-readable recording medium
CN107958091A (en) * 2017-12-28 2018-04-24 北京贝塔智投科技有限公司 A kind of NLP artificial intelligence approaches and interactive system based on financial vertical knowledge mapping
CN108182262A (en) * 2018-01-04 2018-06-19 华侨大学 Intelligent Answer System construction method and system based on deep learning and knowledge mapping
CN109145102A (en) * 2018-09-06 2019-01-04 杭州安恒信息技术股份有限公司 Intelligent answer method and its knowledge mapping system constituting method, device, equipment
CN109388748A (en) * 2018-09-26 2019-02-26 深圳壹账通智能科技有限公司 A kind of answering method of comment information, storage medium and server

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
时雨 ; 古天龙 ; 宾辰忠 ; 孙彦鹏 ; .基于知识图谱的旅游景点问答系统.桂林电子科技大学学报.2018,(04),全文. *

Also Published As

Publication number Publication date
CN110807091A (en) 2020-02-18

Similar Documents

Publication Publication Date Title
CN110807091B (en) Hotel intelligent question-answer recommendation and decision support analysis method and system
Dörr Mapping the field of algorithmic journalism
Liu et al. Analyzing changes in hotel customers’ expectations by trip mode
Law et al. Human computation
CN106709040B (en) Application search method and server
CN104268292B (en) The label Word library updating method of portrait system
CN106663125A (en) Question sentence generation device and computer program
CN109447266B (en) Agricultural scientific and technological service intelligent sorting method based on big data
Ahn Human computation
CN106062730A (en) Systems and methods for actively composing content for use in continuous social communication
CN103679462A (en) Comment data processing method and device and searching method and system
US20160378853A1 (en) Systems and methods for reducing search-ability of problem statement text
CN109906450A (en) For the method and apparatus by similitude association to electronic information ranking
Fawcett et al. Data Science for Business
JP5968744B2 (en) SEARCH METHOD, DEVICE, AND COMPUTER-READABLE RECORDING MEDIUM USING CONCEPT KEYWORD EXTENDED DATA SET
CN112989038B (en) Sentence-level user portrait generation method and device and storage medium
Xue et al. Social media user behavior analysis applied to the fashion and apparel industry in the big data era
Karpagam et al. A mobile based intelligent question answering system for education domain
Antenucci et al. Ringtail: a generalized nowcasting system
Pokrovskii Study on customer behavior analysis using machine learning
Simanjuntak et al. Weighted Hybrid Recommendation System for Toba Tourism Based on Google Review Data
JP7160971B2 (en) content controller
Ennaji et al. Multi-agent framework for social CRM: Extracting and analyzing opinions
Tao et al. Sentiment analysis based on the domain dictionary: A case of analysing online apparel reviews
Jiang et al. Seize market opportunities: market segmentation, profile and monitoring through user-generated content

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
GR01 Patent grant
GR01 Patent grant