Simple data analysis using Weka and excel

Use Weka And Excel Simple data analysis

---

Use Weka Realize linear regression of one variable

example ： A banker wants to calculate the relationship between credit card points and users' monthly income , A file is available bank.arff, The file contains 7 Attributes , Monthly income 、 Working days/month 、 Current credit card limit 、 Historical statistics of the proportion of repayment on time 、 The largest overdraft ever 、 The number of bank loans 、 Credit card points . But the bank only wants to count the relationship between credit card points and monthly income , So when building the model, you need to remove the rest 5 The influence of attributes , Leave only “ Monthly income ” and “ Credit card points ” this 2 Attributes .

This file is a customization file .
bank.arff:

@RELATION creditCardScore
%%%%
%SECTION1:PERSONAL INFO
%%%%
%
% Monthly income 
%
@ATTRIBUTE personInfo.monthlySalary NUMERIC
%%%%
%SECTION2:BUSINESS INFO
%%%%
%
% Working days/month 
%
@ATTRIBUTE businessInfo.workingDayPerMonth NUMERIC
%%%%
%SECTION3:CREDIT CARD INFO
%%%%
%
% Current limit 
%
@ATTRIBUTE creditCardInfo.currentLimit NUMERIC
%
% Monthly normal repayment ratio 
%
@ATTRIBUTE creditCardInfo.percentageOfNormalReturn NUMERIC
%
% The largest overdraft ever 
%
@ATTRIBUTE creditCardInfo.maximumOverpay NUMERIC
%%%%
%
% Number of loans 
%
@ATTRIBUTE financialInfo.personalLoan NUMERIC
%%%%
%RESULT:CREDIT SCORE
%%%%
@ATTRIBUTE creditScore NUMERIC
@DATA
10000,22,20000,1,0,200000,55
15000,20,30000,0.5,14200,20000,78
20000,18,40000,0.6,50000,200000,87
30000,22,60000,0.2,30000,150000,67
22000,15,30000,0.7,20000,140000,71
13200,21,18000,0.9,40000,500000,43
15500,20,30000,0.4,14200,20000,59
25000,26,40000,0.5,50000,200000,88
28670,23,40000,0.7,30000,120000,68
22000,15,40000,0.7,20000,140000,72
10000,18,20000,0.6,30000,150000,47
14300,20,29800,0.5,14200,20000,72
20000,18,40000,0.9,50000,200000,88
34335,22,50000,0.6,30000,150000,74
24555,15,20000,0.9,20000,120000,79
10055,22,80000,1,0,200000,79
15000,20,80000,0.9,90200,20000,86
25400,17,30000,0.7,50000,200000,82
30000,22,70000,0.2,30000,0,72
22000,30,80000,0.7,20000,140000,71

Use Weka Explorer Import data ：

Click on Open file After importing, the result is as follows ：

This is the integrity analysis of the data .Attributes Column display bank.arff Attributes in the file , And clicking each attribute will have a separate analysis . single click Edit Button to view the record of the file .
Such as ：

According to the meaning , Select the unnecessary attribute , single click Remove Remove extra attributes ：

stay Classify Tab, click Choose Button , stay Classifiers Attribute function Under properties, click LineRegression Options for linear regression analysis .

stay Test options The meanings of the options in the area are as follows :

• Use training set： Use all data as model training
• Supplied test set： Set up the test set , After model training , Set up the test data set from here .
• Cross-validation： Divide the data set evenly according to the method of cross validation , Part of it is a training set , Part as a test set
• Percentage split： In proportion , Divide the data set into training set and test set

Choose Use training set Option to experiment , single click Start Button to view the analysis results ：

Parameters in the analysis results ：

• Correlation coefficient： The correlation coefficient
• Mean absolute error： Mean absolute error
• Root mean squared error： Root mean square error
• Relative absolute error： Relative absolute error
• Root relative squared error： Relative square root error
• Total Number of Instance： Number of cases

---

Use Excel Implement polynomial regression

example ： Solve polynomial equations

The existing experimental data are as follows ：

stay file Click... Under the menu Options menu ：

choice add-in , And select Analysis tool library ：

Click on go to , stay Add in Interface selection Analysis tool library Then click OK ：

choice Data analysis , And specify the analysis type as Return to ：

Enter... In the pop-up interface x,y Value input field ：

Click OK to view the analysis results ：