Data Mining And Analysis Of Campus All-in-one Card

With the deepening of education information,the requirements of information technology in Colleges and universities have changed,the concept of smart campus arises at the historic moment,smart campus to text,emphasizing the use of data driven services.Campus card has accumulated a large number of student consumption data and daily behavior data,which can best display the above construction goals and requirements.Therefore,this paper takes this kind of data as the main object of mining analysis,introduces the current research status of data mining of card,expounds the related concepts of data mining technology,and then makes specific work in the following aspects:1.Pretreatment of consumption data and meteorological data of smart card.Including data cleaning,data integration and data completion.2.Exploratory data analysis.Firstly,the descriptive statistical analysis of campus card consumption data,draw a card total consumption of histogram and boxplot,canteens in different time the number of consumer trends and consumption in accordance with the number of draw pictures and words consumer businesses;meteorological data and consumption data were analyzed by Shapiro-Wilk,Test,and that the temperature and precipitation the frequency of consumption are not subject to normal distribution,and then select "Spearman s correlation coefficient correlation verification,by testing the temperature has positive correlation with the frequency of consumption,not related to precipitation and frequency of consumption,then this phenomenon makes further analysis.3.Cluster analysis of student groups.According to the various consumer card data,the use of K-means algorithm for clustering analysis of student groups,according to the within group variance map to determine 5 groups of clustering,while the output clustering centers were clustering4.Student group social network analysis.According to the card dining room consumption data,combined with the concept of social network analysis,draw the student social network diagram,mining social networking talent.5.Construction of recommendation system based on ItemCF algorithm.According to the consumption data of one card dining room,the ItemCF algorithm is used to build a simple recommendation system.First select the users in various consumer businesses as the number of feature vectors,and then calculate the similarity coefficient of Jaccard,get the similarity between different businesses,according to the recommendation system,it can recommend preference businesses for students.Based on campus card data mining and analysis is a wide range,rich content of practical research topics.According to the descriptive statistical analysis of the results,the logistics department can grasp the whole school and business situation,consumption information,decision support for optimization of education resources;according to the correlation analysis of meteorological data and consumption data,can be given in the canteen and takeaway businesses under different weather conditions for meal suggestions;according to cluster analysis and social network analysis the results can be for different people to develop personalized activities for advice,recommend students with similar habits of friends;according to the recommendation system,it can recommend preference businesses for students,not only meet the needs of students,and can increase the business income.Therefore,data mining and statistical analysis technology for campus card data is of great practical significance.