The Design And Implementation Of Combination Analysis And Recommendation Web Application Based On Visual Modeling

We generate data from time to time in our daily work and lives. With the booming development of Internet, we have a variety of access to acquire data, and the data we collect are more and more abundant. In the past people used to leave the data unprocessed after collecting them. But nowadays, people no longer regard data as some kind of old and static record.When a large amount of data are cleaned, processed and analyzed, unexpected patterns and models can be found, which can promote the application and development of information technologies in our daily work and lives, and bring huge economic benefits.The target user group of existing data analysis software and APPs are people with abilities and experiences of data processing, and most of them are not free of charge. Thesoftwareare usually complicated and simply provide the overall optimum recommendation, which are unable to meet common users’ need. Most peoplecannot easily use existing data analysis software to solve practicalproblemsfor they usually lack of professional knowledgeabout data analysis and related algorithms.The system described in this thesis is designed specifically to tackle this problem.Being a web-basedapplicationthe system provides people having no data processing skill with the ability of analyzing data. This system can be generally divided into three function parts, including uploading data files, performing modeling of decision tree, and viewing the analysis report. First, the users must upload their own original data files. They then can construct a model of decision tree. Finally they can look into the result in an analysis report. Apart from the traditional global optimumrecommendation, the system supports users constructing personalized decision tree.The system adopts alightweight front-end design patternMVP and uses backbone.js as the basic framework. We used some open source library such as jQuery, joint.js, handlebars.js etc.in implementation of the system, and performed unit test using mocha.jsto ensure the robustness and validity of system.This paper mainly focuses on describing the author’s work onthe design and implementation of the combination analysis system, including creating user stories, conducting requirement analysis, designing the structure and coding of the system. The paper will also depict the drawbacks of current work and further possible improvements of the system.