英语论文网

留学生硕士论文 英国论文 日语论文 澳洲论文 Turnitin剽窃检测 英语论文发表 留学中国 欧美文学特区 论文寄售中心 论文翻译中心 我要定制

Bussiness ManagementMBAstrategyHuman ResourceMarketingHospitalityE-commerceInternational Tradingproject managementmedia managementLogisticsFinanceAccountingadvertisingLawBusiness LawEducationEconomicsBusiness Reportbusiness planresearch proposal

英语论文题目英语教学英语论文商务英语英语论文格式商务英语翻译广告英语商务英语商务英语教学英语翻译论文英美文学英语语言学文化交流中西方文化差异英语论文范文英语论文开题报告初中英语教学英语论文文献综述英语论文参考文献

ResumeRecommendation LetterMotivation LetterPSapplication letterMBA essayBusiness Letteradmission letter Offer letter

澳大利亚论文英国论文加拿大论文芬兰论文瑞典论文澳洲论文新西兰论文法国论文香港论文挪威论文美国论文泰国论文马来西亚论文台湾论文新加坡论文荷兰论文南非论文西班牙论文爱尔兰论文

小学英语教学初中英语教学英语语法高中英语教学大学英语教学听力口语英语阅读英语词汇学英语素质教育英语教育毕业英语教学法

英语论文开题报告英语毕业论文写作指导英语论文写作笔记handbook英语论文提纲英语论文参考文献英语论文文献综述Research Proposal代写留学论文代写留学作业代写Essay论文英语摘要英语论文任务书英语论文格式专业名词turnitin抄袭检查

temcet听力雅思考试托福考试GMATGRE职称英语理工卫生职称英语综合职称英语职称英语

经贸英语论文题目旅游英语论文题目大学英语论文题目中学英语论文题目小学英语论文题目英语文学论文题目英语教学论文题目英语语言学论文题目委婉语论文题目商务英语论文题目最新英语论文题目英语翻译论文题目英语跨文化论文题目

日本文学日本语言学商务日语日本历史日本经济怎样写日语论文日语论文写作格式日语教学日本社会文化日语开题报告日语论文选题

职称英语理工完形填空历年试题模拟试题补全短文概括大意词汇指导阅读理解例题习题卫生职称英语词汇指导完形填空概括大意历年试题阅读理解补全短文模拟试题例题习题综合职称英语完形填空历年试题模拟试题例题习题词汇指导阅读理解补全短文概括大意

商务英语翻译论文广告英语商务英语商务英语教学

无忧论文网

联系方式

using Excel:Credit Scoring and Data Mining [2]

论文作者:anter论文属性:建议书 proposal登出时间:2010-12-03编辑:anterran点击率:6594

论文字数:3245论文编号:org201012031045486098语种:英语 English地区:中国价格:免费论文

关键词:Credit ScoringData Mining Computer Lab Workshopusing Excel

ables decide what would be good/bad ratio for interactions of
cheque account
deposit account
current account
phone
Can you work out more than two interactions?

5. For continuous variables do same thing and then combine to get sensible coarse classifying for age
Why does pivot tables not work as well for this.
Can you construct the best monotone coarse classifier?


B: Constructing KS and ROC Curves

Suppose one has constructed a scorecard how does one then calculate the KS and ROC curves?
We will use the linear regression based scorecard we obtained in the previous workshop where

100Y= -1 +17X(18-33) -11X(33-48) -31X(49-54) +19X(55 -63) + 68X(0child) +63X(1child) +67X(2child) +30X(B,Y) +6X(R,H,X) –14X(Z, no response) +23X(P,E,L)

Go to sheet 2 where the data used in constructing this scorecard is. In row 104 type in the scoring values starting in column E. So type in 17 in E104, -11 in F104 , -31 in G104 and so in until one finishes with 23 in O104.
Go to T1 and type in “score”.
In T2 we calculate the score of that person by typing in =SUMPRODUCT(E2:O2,E$104:O$104)
This should give a score of 3.
Copy cell T2 down the column to T101, che
Copy column D, the good=1 column into column U by highlighting the top of column D ( the cell with a D in it , clicking copy and then going to the top of U ( the cell with U) in it and paste.

These, score and whether person was good are the two pieces of information needed to evaluate the ROC and KS curve. We will use another sheet to evaluate the curves on.
Highlight columns T and U. Press copy go to Sheet 4 and with cursor on the first column go to Paste Special in the Edit menu. Click the values box ( otherwise we only get the formulae which don’t make sense in this sheet) and then click on OK.
Highlighting the two columns A and B go to Sort in the Data menu and sort on score column with scores ascending.
In cell B102 ( at bottom of good column) use Summation sign to get =SUM(B2:B101), the number of goods, In c102 type =100-b102 and in d102 type =b102. We will use the C and D column to find how many bads ( C) and How many goods(D) have values less than that customer.
Type 0s in C1 and D1
In C2, type =IF(b2=0,C1+1,C1) which adds one if the b column is 0 ( a bad). Copy this down to C101.
In D2 type =IF(b2=1,d1+1,D1) which does the same thing for the goods.
To get these into percentages in E1 type %of bads, in F1 type % of goods
In E2 type =C2/C$102 and copy down the E column.
In F2 type =D2/D$102 and copy down the F column

To get KS value in g column type in G2 =E2-F2 and copy this down the whole column. The largest value ( 0.66) is the KS statistic

To get KS chart go to Chart Wizard
Choose line and in data range type in or highlight E1 to F101 ( E1:F101). Let this chart be in the current sheet and tidy up the colours and background fill. This is the KS chart

留学生论文https://www.51lunwen.org/StudentPapers.htmlTo get the ROC chart go to Chart Wizard againChoose scatter this time and choose the variant where the data points are connected by a line, The data range is E2:F101 and when it comes to labelling the axes label the x axis % bads and the Y axis % goods.

Exercise: Can you work out how to calculate the Gini coefficient form the numbers given in the spreadsheet?

论文英语论文网提供整理,提供论文代写英语论文代写代写论文代写英语论文代写留学生论文代写英文论文留学生论文代写相关核心关键词搜索。

共 2/2 页首页上一页12下一页尾页

英国英国 澳大利亚澳大利亚 美国美国 加拿大加拿大 新西兰新西兰 新加坡新加坡 香港香港 日本日本 韩国韩国 法国法国 德国德国 爱尔兰爱尔兰 瑞士瑞士 荷兰荷兰 俄罗斯俄罗斯 西班牙西班牙 马来西亚马来西亚 南非南非