Research Of Rough Set And Granular Computing Theory Application In Air Tickets Recommender System

The recommendation system is becoming a specialized field of study with the promotion of the huge economic benefits by electronic commerce. The programming and application are the two main components of a recommendation algorithm and a well system is an intelligent recommendation service system for some specific areas and the robustness of a recommendation algorithm is just reflected some specific data sets that is a metrics of the algorithm. With the diversification and diverse information of an e-commerce platform that brings a massive growth in the scale and numerous forms of information resources. The traditional recommendation algorithm exhibits a series of questions such as gradually reveal worse recommendation, slowly information processing speed through the processing of that massive information. Introducing the new knowledge and theories to improve those algorithms or program the new recommendation algorithm is becoming more and more significant.An air ticket e-commerce platform is different from an ordinary e-commerce platform. The first of the most, those discrepancies come from the special attributes of the air ticket that decided the traditional recommendation algorithm couldn’t handle with timeliness and significance. Secondly, in order to increase ticket sales, most ticketing portal use a direct purchase pattern that cannot register the platform account and without authenticating identity, user just to provide a valid mobile number and identity card to buy an air ticket. This pattern brings great extents that promote the tickets sales and convenient for users. But that also brings the difficult to obtain a user’s truthful information and so on. To resolve these problems, the traditional recommendation technologies are difficult to achieve the precise recommendation application.Improving the technical shortcomings of traditional recommendation with application of the airline ticketing recommended and perfecting the current recommendation algorithm of the mainstream air-ticketing platform to adapt the processing of the massive tickets information. We use the Rough set theory technologies and granular computing theories to deal with the air ticketing data that analyzing by the original the discrete data, attribute reduction and value reduction, etc...Proposed a discrete method to discrete data by using the importance of the breakpoint and attribute reduction method attribute importance for which knowledge discovery, then extracting the preliminary decision rules. Dealing with incremental feature of ticketing data, by using the resolution characteristics of the matrix and multi-granularity principle that to analysis and process the incremental data to extract new decision rules. At last, extracting the completion decision rules of the extraction ticket data sets.Through the analyzing of the air ticketing raw data sets, extracting appropriate decision-making rules and programming the air tickets recommendation algorithm based on knowledge discovery. Providing a guideline for major ticketing platform to design a new efficient data processing and program a new air tickets recommendation algorithm. While improving and perfecting the short of the domestic air ticketing recommendation system in this regard at present. The main work and the results of the study as follows:1. Analyzing the data of air ticket from the Quare with using the vertical search rule of the website is combined with the web crawler algorithm based on depth principle to extract aviation ticket data;2. Researching the application of Rough set in data preprocessing, and analyzing the Rough Set Discretization, then given the cognizable matrix row-discrete algorithm using the degree of importance on the breakpoint;3. Researching the method of attribute reduction and value reduction of decision table then given the initial decision rules on the air ticketing data;4. The rules of incremental data acquisition through the analysis of the improved discernibility matrix in the rules for incremental data acquisition, incremental data, and the dynamic access to new decision rule algorithm and, completed the final decision rule set;5. Based on the final decision rule set and designed the recommendation algorithm of aviation ticket that based on knowledge discovery.