CN-121998138-A - Scenic spot ticket stock and intelligent question-answer linkage management method and system

CN121998138ACN 121998138 ACN121998138 ACN 121998138ACN-121998138-A

Abstract

The invention discloses a management method and a system for scenic spot ticket stock and intelligent question-answer linkage, which relate to the technical field of data processing, and are used for receiving a user query statement, dividing the user query statement into fragment sets with semantic identifications, matching preset training semantic fragments to obtain candidate fragments, generating a semantic decoding sequence after hierarchical analysis and constructing thinking chain guide information, extracting key elements through a large language model by combining the information and target fragments, generating and outputting a verification path set and a ticket scheme set through a business rule network, and responding to user selection to execute stock linkage operation. The method simplifies semantic analysis by guiding information through a thinking chain, strengthens interpretability, accurately extracts ticket purchasing elements, realizes continuous response from inquiry to ticket scheme and inventory operation through a full-flow closed loop, breaks out island of consultation and inventory information, optimizes ticket purchasing links, ensures real-time and accurate scheduling and improves service accuracy.

Inventors

Li Shaokuo
WEI SHAOKANG
LU HAO

Assignees

宋城独木桥网络有限公司

Dates

Publication Date: 20260508
Application Date: 20260119

Claims (10)

1. A method for managing scenic spot ticket stock and intelligent question-answer linkage, the method comprising: Receiving a query statement input by a user; dividing the query statement into a plurality of semantic segments to obtain a semantic segment set, wherein the semantic segments in the semantic segment set comprise semantic type identifiers and text segments corresponding to the semantic type identifiers; Matching and searching a preset training semantic segment set according to a target semantic segment, and screening semantic segments with highest similarity with the target semantic segment to obtain candidate semantic segments; Performing hierarchical traversal analysis on the candidate semantic segments to obtain semantic decoding sequences; The method comprises the steps of constructing thinking chain guide information according to the semantic decoding sequence, wherein the thinking chain guide information is a semantic analysis guide method for decomposing a query statement into local semantic fragments, inputting the local semantic fragments into a preset large language model according to priority for processing, and the guide information is used for reducing the complexity of semantic analysis so as to enable the preset large language model to focus on the recognition and combination of the local semantic fragments; Inputting the target semantic segment and the thinking chain guide information into a preset large language model to decode the semantic segment to obtain key elements; Counting key elements of all semantic segments to obtain a key element set, and generating a plurality of initial verification paths in a preset business rule network according to the key element set to obtain an initial verification path set; generating a ticket scheme set according to the initial verification path set; And outputting the ticket scheme set to a user interface, responding to a selection instruction of a user for a specific alternative ticket scheme, and executing inventory linkage operation based on a verification path corresponding to the scheme.
2. The method for managing scenic spot ticket stock and intelligent question-answer linkage according to claim 1, wherein the step of performing matching search on a preset training semantic segment set according to a target semantic segment, and the step of screening semantic segments with highest similarity to the target semantic segment to obtain candidate semantic segments comprises the following steps: extracting semantic type identifiers in the target semantic segments to obtain a semantic type identifier set; Screening all training semantic segments containing the same semantic type identification from a preset training semantic segment set according to the semantic type identification set to obtain an initial training semantic segment set; Extracting semantic type identifiers of target initial training semantic segments to obtain an initial semantic type identifier set, wherein the target initial training semantic segments are any one initial training semantic segment in the initial training semantic segment set; Calculating Jaccard similarity between the semantic type identification set and the initial semantic type identification set to obtain a first similarity value; Calculating the semantic similarity of the text segment corresponding to the target semantic segment and the training text segment corresponding to the target initial training semantic segment to obtain a second similarity value; Weighting the first similarity value and the second similarity value to obtain a comprehensive similarity value; and counting the comprehensive similarity values of all initial training semantic segments and the target semantic segments, and selecting the training semantic segment with the highest comprehensive similarity value as a candidate semantic segment.
3. The method for managing scenic spot ticket inventory and intelligent question-answering linkage according to claim 1, wherein constructing the thinking chain guide information according to the semantic decoding sequence comprises: Constructing a semantic tree according to the semantic decoding sequence, extracting non-terminal nodes of the semantic tree as primary semantic segments, filtering redundant segments in the candidate semantic segments, and performing hierarchical traversal analysis on the filtered primary semantic segments through a preset target traversal algorithm to obtain a hierarchical semantic sequence; Identifying semantic type identifiers of the initially selected semantic segments in the hierarchical semantic sequence, association relations of text segments and traversal sequences according to the hierarchical relation of the semantic tree to obtain sequence structure feature information; according to the structural feature information, the semantic decoding sequence is disassembled into ordered steps to obtain an ordered step set; And combining the ordered step sets according to the traversing order in the sequence structural feature information to obtain the thinking chain guide information.
4. The method for managing scenic spot ticket stock and intelligent question-answer linkage according to claim 1, wherein inputting the target semantic segment and the thinking chain guide information into a preset large language model to perform semantic segment decoding to obtain a key element set comprises: Analyzing the hierarchical association rule and the priority in the semantic decoding sequence corresponding to the target semantic segment according to the thinking chain guide information to obtain semantic analysis information; Determining the analysis sequence of the target semantic segment according to the semantic analysis information, and decomposing the analysis task of the target semantic segment into a plurality of semantic analysis subtasks which are sequentially executed according to the analysis sequence to obtain a subtask set; Performing token coding on the target semantic segment to obtain a coding sequence; Acquiring context semantic information corresponding to the target semantic segment; In the target subtasks, screening effective token according with the corresponding subtask corresponding level semantic rule from the coding sequence according to the context semantic information to obtain a candidate token set; carrying out semantic combination and relation construction on the candidate token set according to the sequence of core semantics superior to subordinate semantics to obtain hierarchical semantic information; And extracting key information and removing redundant information from the hierarchical semantic information of all subtasks to obtain key elements.
5. The method of claim 1, wherein generating a ticket schema set from the initial validation path set comprises: constructing a verification state matrix according to the verification path set, wherein the rows of the verification state matrix correspond to initial verification paths and the columns correspond to verification nodes; for each verification node in the verification state matrix, corresponding inventory state data is obtained in real time, and the inventory state data is filled into corresponding positions of the matrix to obtain the inventory state matrix; Calculating the feasibility degree score of each initial verification path in the initial verification path set according to the inventory state matrix to obtain a feasibility degree score set; Selecting an initial verification path with the feasibility degree score larger than a preset threshold value in the feasibility degree score set as an effective verification path to obtain an effective verification path set; And generating a ticket scheme for each effective verification path in the effective verification path set based on the constraint conditions of each verification node and real-time inventory data, and finally obtaining a ticket scheme set.
6. A scenic spot ticket inventory and intelligent question-answer linkage management system, the system comprising: The query receiving module is used for receiving a query statement input by a user; The dividing module is used for dividing the query statement into a plurality of semantic segments to obtain a semantic segment set, wherein the semantic segments in the semantic segment set comprise semantic type identifiers and text segments corresponding to the semantic type identifiers; The screening module is used for carrying out matching search on a preset training semantic fragment set according to a target semantic fragment, and screening semantic fragments with highest similarity with the target semantic fragment to obtain candidate semantic fragments; The hierarchical analysis module is used for performing hierarchical traversal analysis on the candidate semantic segments to obtain a semantic decoding sequence; The system comprises a semantic analysis and guide module, a semantic analysis and guide module and a semantic analysis and guide module, wherein the semantic analysis and guide module is used for constructing the semantic analysis and guide information according to the semantic decoding sequence, wherein the semantic analysis and guide information is a semantic analysis and guide method for decomposing a query statement into local semantic fragments and inputting the local semantic fragments into a preset large language model for processing according to priority, and the guide information is used for reducing the complexity of semantic analysis so as to focus the preset large language model on the identification and combination of the local semantic fragments; The decoding module is used for inputting the target semantic segment and the thinking chain guide information into a preset large language model to decode the semantic segment to obtain key elements; the statistics module is used for counting key elements of all semantic segments to obtain a key element set, and generating a plurality of initial verification paths in a preset business rule network according to the key element set to obtain an initial verification path set; The scheme generation module is used for generating a ticket scheme set according to the initial verification path set; and the inventory linkage module is used for outputting the ticket scheme set to a user interface, responding to a selection instruction of a user for a specific alternative ticket scheme, and executing inventory linkage operation based on a verification path corresponding to the scheme.
7. The scenic spot ticket inventory and intelligent question-answering linked management system of claim 6, wherein the screening module comprises: The first extraction module is used for extracting semantic type identifiers in the target semantic segments to obtain a semantic type identifier set; The first screening module is used for screening all training semantic segments containing the same semantic type identification from a preset training semantic segment set according to the semantic type identification set to obtain an initial training semantic segment set; the second extraction module is used for extracting semantic type identifiers of target initial training semantic segments to obtain an initial semantic type identifier set, wherein the target initial training semantic segments are any initial training semantic segment in the initial training semantic segment set; The first calculation module is used for calculating the Jaccard similarity between the semantic type identification set and the initial semantic type identification set to obtain a first similarity value; The second calculation module is used for calculating the semantic similarity between the text segment corresponding to the target semantic segment and the training text segment corresponding to the target initial training semantic segment to obtain a second similarity value; the weighting module is used for weighting the first similarity value and the second similarity value to obtain a comprehensive similarity value; The candidate semantic segment generation module is used for counting the comprehensive similarity values of all initial training semantic segments and the target semantic segments, and selecting the training semantic segment with the highest comprehensive similarity value as the candidate semantic segment.
8. The scenic spot ticket inventory and intelligent question-answer linked management system of claim 6, wherein the thinking chain information construction module comprises: The hierarchical semantic generation module is used for constructing a semantic tree according to the semantic decoding sequence, extracting non-terminal nodes of the semantic tree as primary semantic segments, filtering redundant segments in the candidate semantic segments, and performing hierarchical traversal analysis on the filtered primary semantic segments through a preset target traversal algorithm to obtain a hierarchical semantic sequence; the identification module is used for identifying semantic type identification of each initially selected semantic segment in the hierarchical semantic sequence, association relation of text segments and traversal order according to the hierarchical relation of the semantic tree to obtain sequence structure feature information; the step disassembling module is used for disassembling the semantic decoding sequence into ordered steps according to the structural feature information to obtain an ordered step set; And the combination module is used for combining the ordered step set according to the traversing order in the sequence structural feature information to obtain the thinking chain guide information.
9. The scenic spot ticket inventory and intelligent question-answering linked management system of claim 6, wherein the decoding module comprises: the semantic information generation module is used for analyzing the hierarchical association rule and the priority in the semantic decoding sequence corresponding to the target semantic segment according to the thinking chain guide information to obtain semantic analysis information; the subtask generation module is used for determining the analysis sequence of the target semantic segment according to the semantic analysis information, and decomposing the analysis task of the target semantic segment into a plurality of sequentially executed semantic analysis subtasks according to the analysis sequence to obtain a subtask set; the coding module is used for performing token coding on the target semantic segment to obtain a coding sequence; the information acquisition module is used for acquiring context semantic information corresponding to the target semantic segment; The second screening module is used for screening effective token conforming to the corresponding hierarchical semantic rule of the corresponding subtask from the coding sequence according to the context semantic information in a target subtask to obtain a candidate token set; the semantic information generation module is used for carrying out semantic combination and relation construction on the candidate token set according to the sequence of the core semantic better than the subordinate semantic to obtain hierarchical semantic information; And the key element generation module is used for extracting key information and removing redundant information from the hierarchical semantic information of all subtasks to obtain key elements.
10. The scenic spot ticket inventory and intelligent question-answer linked management system of claim 6, wherein the scheme generation module comprises: The matrix construction module is used for constructing a verification state matrix according to the verification path set, wherein the rows of the verification state matrix correspond to the initial verification paths, and the columns correspond to the verification nodes; The inventory state matrix generation module is used for acquiring corresponding inventory state data in real time for each verification node in the verification state matrix, and filling the inventory state data to the corresponding position of the matrix to obtain an inventory state matrix; the scoring module is used for calculating the feasibility score of each initial verification path in the initial verification path set according to the inventory state matrix to obtain a feasibility score set, wherein the feasibility score is calculated based on the verification node proportion meeting the constraint in the path and the key constraint meeting state; the effective path screening module is used for selecting an initial verification path with the feasibility degree score larger than a preset threshold value in the feasibility degree score set as an effective verification path to obtain an effective verification path set; And the ticket scheme generating module is used for generating a ticket scheme for each effective verification path in the effective verification path set based on the constraint condition of each verification node and the real-time inventory data, and finally obtaining a ticket scheme set.

Description

Scenic spot ticket stock and intelligent question-answer linkage management method and system Technical Field The invention belongs to the technical field of data processing, and particularly relates to a method and a system for managing scenic spot ticket stock and intelligent question-answer linkage. Background With the digital conversion promotion of the text industry, scenic spot ticket management is updated from the traditional manual ticket selling and ticket checking mode to an intelligent mode led by an intelligent ticket management system, through integrating OTA (over-the-air) platforms, self-contained small programs, off-line windows and other multi-channel resources, real-time synchronization of orders, stock dynamic management and multi-mode quick checking are realized, a part of systems are integrated with big data analysis functions, order sales, order withdrawal conditions and tourist portraits can be counted in real time, meanwhile, the connection of ticket business, accommodation, catering and other multi-business scenes is supported, and automation equipment such as self-service ticket vending machines and intelligent gate machines is matched, so that the scenic spot operation efficiency is remarkably improved, the labor cost is reduced, and the scenic spot is promoted to be converted from flow dependence to value creation. However, the conventional scenic spot ticketing system has obvious defects in the analysis of user sentences, and an intelligent question-answering module can only perform shallow analysis on general sentences such as scenic spot profile, open time and the like, so that the user query sentences cannot be split into accurate fragments with semantic identifications, the hierarchical analysis capability of the sentences is also lacking, the inherent logic of semantic structures is difficult to comb, the semantic analysis complexity is high, the interpretability is poor, key elements related to ticket purchase cannot be accurately extracted from the user sentences by the conventional system, the sentence analysis result cannot be linked with ticket stock management, so that sentences of ticket core requirements such as residual tickets, reservation names and the like of users cannot be accurately responded, and information islands of consultation and stock management are formed, and the response accuracy of ticket service is low. Disclosure of Invention The invention aims to solve the problems that the conventional scenic spot ticket business system has insufficient analysis capability on user sentences, and analysis results cannot be effectively linked with ticket business inventory management, so that the response accuracy of ticket business service is low, thereby providing a scenic spot ticket inventory and intelligent question-answer linked management method and system. In a first aspect of the present invention, a method for managing scenic spot ticket stock and intelligent question-answer linkage is provided, where the method includes: Receiving a query statement input by a user; dividing the query statement into a plurality of semantic segments to obtain a semantic segment set, wherein the semantic segments in the semantic segment set comprise semantic type identifiers and text segments corresponding to the semantic type identifiers; Matching and searching a preset training semantic segment set according to a target semantic segment, and screening semantic segments with highest similarity with the target semantic segment to obtain candidate semantic segments; Performing hierarchical traversal analysis on the candidate semantic segments to obtain semantic decoding sequences; constructing thinking chain guide information according to the semantic decoding sequence, wherein the thinking chain guide information is a semantic analysis guide method for decomposing a query statement into local semantic segments, inputting the local semantic segments into a preset large language model according to priority for processing, and the guide information is used for reducing the complexity of semantic analysis so as to focus the preset large language model on the identification and combination of the local semantic segments; Inputting the target semantic segment and the thinking chain guide information into a preset large language model to decode the semantic segment to obtain key elements; Counting key elements of all semantic segments to obtain a key element set, and generating a plurality of initial verification paths in a preset business rule network according to the key element set to obtain an initial verification path set; generating a ticket scheme set according to the initial verification path set; And outputting the ticket scheme set to a user interface, responding to a selection instruction of a user for a specific alternative ticket scheme, and executing inventory linkage operation based on a verification path corresponding to the scheme. According to the scheme, the