Research And Application Of Web Visualization Technology In Data Mining

Recently,with the popularity of the rapid development in computer hardware and software,as well as the internet,the amount of data from the internet and mobile devices grows rapidly.It is important to analyze the useful information from the massive data.Current development of internet technology provides an effective way to explore this issue.Data mining is an effective way,because it can extracts potential and valuable information from a large number of data sets.In addition,visualization is the process of transforming data and information features into visual expressions.In the theoretical research and application of data mining,the combination with visualization technology has produced another important research direction,i.e.,visual data mining.Through the combination of visual technology and human visual characteristics,the information generated by data mining is presented to the user in an intuitive way,making the mining results more valuable and easier to be understood.Most of the data generated by network information is generated through network access and network user behavior.Most of them are stored in large-scale resource database of network information platform.This data is often variable in fields and have different data format,which can be text,image,voice,video etc.The analysis and application of these database log files and data files are the key to the analysis of conversion from e-commerce data flow to information flow.In this thesis,firstly,we introduce the background of large data.Then,the web visualization and data mining technology,as well as the research status and research significance of visualization technology in data mining are introduced.Secondly,we discuss the basic technology of web visualization and data mining.The basic flow of visualization,the common Web front-end visualization tool library,the multi-dimensional scale algorithm and Hadoop distributed processing system are introduced.It provides a technical support for the overall design.Finally,we show several examples for the application of visualization technology in data mining.For the general multidimensional attribute variable data,we use multidimensional scaling algorithm to simplify the research object of multidimensional space into low-dimensional space for locating and analyzing.The characteristics of information data and the results of visualization of the data are given.For the log data of a business forum,Hadoop massive data processing system,combined with distributedfile system(HDFS)and Map / Reduce distributed computational model are used.The data model are built from many aspects,such as the number of registered users,the number of independent IP,the number of jumps and so on.We construct the visual data model,and design the overall technical scheme included in the process from the background data mining to the front visualization display.