Bayesian probability problems separator analysis scholarship

1Bai Rui

Shanghai Second Polytechnic University, College of Computer and Information, e-commerce system,10-A2 class

1248750477@qq.com

Abstract

Our group has been discussed for some time, finalized separator using Bayesian probability of scholarships for data mining. In this process, I participated in the panel discussion topics and design ideas, which also includes the use of Bayesian probability formula scholarships, data entry. Analysis of data: the probability of the first two classes scholarship count out, and then use the Bayesian formula calculates the probability of obtaining a scholarship from which classes the students, the final conclusions: a randomly selected students from two classes, it is known the students are scholarship, students from the possibility of this 10 meter class A1 biggest supplier. The number of students often go to the library, usually late for class times, before and after class seat, the contest number, dormitories score three times and other data clustering analysis, the final conclusion: The first category is the most likely to get a such as scholarships and above, the second category is more likely to get a non-first-class and scholarship, and the third category is those who can not get a scholarship.

Keywords: Bayesian formula1, Scholarship, Cluster analysis

1. Introduction

With the college enrollment, a substantial increase in the number of students, to college students management, teaching work to bring a severe test. The traditional teaching methods can not meet the management and rapid development universities. Many colleges and universities at this stage on student achievement, student basic information is still stuck in the traditional, simple database management and query stage, you can not play its due role. Take an example of student achievement, teacher knowledge on student achievement to make a simple excellent, good, fair, poor assessment does not consider the factors that influence student achievement, some may be subjective factors, some may be objective factors. If certain objective factors such as the learning environment, teachers, etc. can not be a good solution, it will seriously affect student achievement, restricting the development of students, but also seriously hindered the pace of development of school education. Therefore, analysis of student achievement and other key information, such as data mining by technical rationality, improve teaching quality and level of teachers and students is one of the issues most concern.

2. Bayesian analysis of the probability of getting the scholarship formula

The total number two classes of 75, 38 scholarship winners, which accounted for scholarship class A1 ratio of 58%, A2 class winners representing the proportion of the total number scholarships to 42%.

X indicates that a student is selected scholarship winners

Y = i said selected students from class i.

P (X | Y = 10-A1) = 0.56, P (X | Y = 10-A2) = 0.44

P(Y=10-A1)=58%，P(Y=10-A2)=42%

P(X)=P(X|Y=10-A1)P(Y=10-A1)+P(X|Y=10-A2)P(Y=10-A2)

=0.56*0.58+0.44*0.42=0.3248+0.1848=0.5096

So, then choose from two classes a student scholarship in the probability is 0.5096.

According to Bayes' theorem can be obtained:

Conclusion: The two classes were randomly selected from a student, the student is known to scholarship, this 10 meter supplier students from class A1 most likely.

3.Cluster analysis

x1: one semester to the library number (seven days a week by calculating maximum 112 times)

x2: one semester late times (according to four days per week basis, limit 64 times)

x3: one semester in the front seat number (according to four days per week basis, limit 64 times)

x4: one semester to participate in various competitions times (5 times the upper limit of each semester)

x5: one semester dorm Average Rating (limit 20 minutes)

：96-112；：0-2；：60-64；：3-5；：19.5-20

Conclusion:

The first category is the most likely to get the first-class and above scholarship, the second category is more likely to get a non-first-class and scholarship, and the third category is those who can not get a scholarship

4. Semester summary

Through this semester for business intelligence this course, I learned how to target data mining and analysis using Bayesian classifier problems. This semester also includes a number of related technologies and business intelligence software to use to make themselves benefited, so a lot of my rich business intelligence related technical knowledge, but also improved my ability to operate the software, the use of data between link, you can find hidden information and data outside of, for me, this ability is critical to society in the future, people will not have a skill, increase their competitiveness. Thank closely with team members, so that we can successfully complete the task of the working group.

5 .References

[1] Caitlin N. Rinke，Mary R. Williams, Discriminant analysis in the presence of interferences: Combined application of target factor analysis and a Bayesian soft-classifier, Analytica chimica acta, vol. 753, 2012.

[2] Jingxue Yang, Classification of 10 m-resolution SPOT data using a combined Bayesian Network Classifier-shape adaptive neighborhood method，ISPRS journal of photogrammetry and remote sensing , 72(Aug.), 2012.

[3] A. J. M. Abu AfzaDewan Md. FaridChowdhury Mofizur Rahman, World of Computer Science and Information Technology Journal, USA 1(3), 2012

个人作业-商务智能-用贝叶斯分离器分析奖学金概率问题