$1.$ Introduction

You are given two real datasets which contain individual information in US$.$ In the

first real dataset, each individual is associated with $13$ attributes and $1$ additional

Boolean attribute called $$income$$ indicating whether the individual had an income $>$

$50$K $($per year$)$ or not. The second real dataset is the same as the first dataset but the

second one contains only the first $13$ attributes but no attribute $$income$.$ The

objective of this project is to predict whether each individual in the second dataset has

an income $>50$K or not.

There are three phases in this project $$ Phase $1,$ Phase $2$ and Phase $3.$ In Phase $1,$ you

are required to generate an Excel file from two raw files together with attribute names.

In Phase $2,$ you are required to write a design report for this project. In Phase $3,$ you

are required to follow the design report in Phase $2,$ generate the predicted attribute

files for the second real dataset and write a final report

$2.$ Milestones

$1.$ Phase $1$

i$.$ You are given two real datasets in TEXT format, $$training$.$txt$$ and $$test$.$txt$.$

ii$.$ File $$training$.$txt$$ contains $13$ attributes and $1$ additional Boolean attribute.

iii. File $$test$.$txt$$ contains $13$ attributes only.

iv$.$ Open these two TEXT files with MS Excel

v$.$Save them in one MS Excel file where the content of $$training$.$txt$$ is included in $$Sheet $1$ and the content of $$test$.$txt$$ is included in $$Sheet

$2.$ Please re$-$name $$Sheet $1$ as $$training$$ and re$-$name $$Sheet $2$ as $$test$$ in MS Excel.

vi$.$Insert a row at the beginning of each of the two sheets of the Excel file where this row gives the attribute names specified in Section $3$

$2.$Phase $2$

i$.$ In Phase $2,$ you are required to write a design report for this project.

ii$.$You should list $5$ possible data mining models you want to try

iii.Note that a possible data mining model can be $$Decision Tree Classifier$$ with a set of parameters and another possible data mining model can be $$Decision Tree Classifier$$ with another set of parameters. Obviously, one of the possible data mining models can be $$Nearest Neighbor Classifier$$ with a set of parameters.

$3.$

Phase $3$

I. In Phase $3,$ you are required to follow the design report in Phase $2,$ use the XLMiner software to predict attribute $$income$$ for the second real dataset and write a final report.

ii$.$ The final report should include the following.

a$)$ All materials in your design report written in Phase $2($i$.$e$.,$ the $5$ possible data mining models$)$

b$)$ Description of the XLMiner results for each of $5$ possible data mining models

c$)$ Two examples illustrating what attributes determine an individual to have an income $>50$K or not for each of $5$ possible data mining models.

d$)$ Conclusions drawn from each of $5$ possible data mining models and an overall conclusion

iii. In addition to the final report, you are required to generate $5$ predicted attribute files for the second real dataset in TEXT file format. Note that each predicted attribute file corresponds to the output of a possible data mining model you proposed in Phase $2.$ The file format is described in

Section $4.$

$4.$ File format of Predicted Attribute File

In Phase $3,$ you are required to submit $5$ predicted attribute files for the second dataset.

The files should be named as $$predicted$1.$txt$,$predicted$2.$txt$,$predicted$3.$txt$,$

$$predicted$4.$txt$$ and $$predicted$5.$txt$$ where $$predicted$1.$txt$$ corresponds to the output

of the first data mining model proposed in Phase $2$ and the other files have a similar

meaning. The file format of each file is shown as follows.

$<mn>1</mn>$

st

row: $1$ or $0$ where $1$ corresponds to that the first individual in the second dataset

has an income $>50$K and $0$ corresponds to that s$/$he does not$>$

$<mn>2</mn>$

nd

row: $1$ or $0$ where $1$ corresponds to that the second individual in the second

dataset has an income $>50$K and $0$ corresponds to that s$/$he does not $>$

$...$

Here is a sample file.

$1$

$0$

$1$

$0$

$0$

$0$

$0$

$0$

$...$

We have an answer file for the predicted attribute file. Among $5$ files given by you,

we will select the one with the highest accuracy as the final file for marking.Data Specifications

There are $13$ attributes in the first dataset and $1$ additional Boolean attribute called

"income".