Design And Implementation Of College Financial Consultation Question Answering Prototype System Based On NLP

The financial management of colleges has a very strong particularity,which not only follows the basic financial management system,but also conforms to the regulations of the Ministry of education and various scientific research departments,involving many rules.At the same time,the existing software systems for college finance are mostly about college finance business processing,lacking the content query function related to the question and answer of college finance.And in the Internet era,people are more and more inclined to query information on the Internet,but the search engine model requires users to further filter according to the feedback answers,which may not directly obtain the relevant information;the content in the financial field of colleges has certain particularity,so it is difficult to obtain completely accurate information on the Internet.Therefore,this thesis aims at the field of college finance,integrates the professional question and answer data resources of the college finance through the system,and uses natural language processing related technologies to understand the questions asked by the user,complete the precise answers,so as to realize the transfer from traditional "offline consultation" to "robot dialogue".In view of the above background,this thesis designs a prototype system based on natural language processing technology,which can realize the management of related Q & A pairs in the financial field of colleges,provide high-precision answers according to the questions asked by users,and provide Q & A services for users.This thesis mainly studies the following: firstly,through system requirements analysis,this thesis analyzes the business model of college financial consulting,clarifies the business functional requirements,business processes and data processes.Secondly,the detailed design of the business module of the college financial consultation Q & A system,including the system overall design,function module detailed design,question and answer model design and database design.The designed question answering system of college financial consultation mainly includes five functions: standard question management,financial knowledge management,multi round dialogue management,question and answer consultation and feedback,question optimization management.The designed question and answer model include data processing,dialogue intention recognition,multi round dialogue,question and answer matching and answer reply.Then,in the model algorithm implementation stage,specific algorithms are implemented for the designed question and answer model.According to natural language processing related technologies,Chinese word segmentation and word vector representation are completed,so as to more accurately understand the user's language or question,and improve the effectiveness of dialogue with the user.In the process of question and answer matching,the similarity matching algorithm is used to complete the high similarity matching between the question asked and the existing standard question,and the feasibility and accuracy of the algorithm are verified through experiments.Finally,the prototype interface of the system is displayed,including the mobile terminal for user questions,and the background management terminal for college financial staff manage question and answer information.The question and answer system of college financial consultation researched in this thesis can not only capture the questioner's intention of asking questions efficiently and provide standard answers with high matching degree,but also support the continuous optimization and improvement of the consultation system in the process of providing consultation,and improve the degree of college financial domain informatization in the aspect of consultation and answer,so as to provide reference for the development of college financial consultation informatization.