Association Rule Evaluation

In this page you can see how the association rules can be extracted and evaluated from a dataset which contains information about the number of rents by days with additional information such as weather conditions and the characteristics of the day. The target variable is the total number of rents per day (cnt). You will use the data file “bike.csv”.

You can find the report file that contains also the necessary  R code by clicking --> here

Mining on KDD CUP 2009 up-selling data.

Here you can find our mining study on KDD CUP 2009 - Customer Relationship Prediction. The data we choose is up-selling data. You can find the needed information from http://www.kdd.org/kdd-cup/view/kdd-cup-2009 . In this project R Studio is used for analysing the data.

Press here for our report!

The R codes can be found below as txt files (You can directly change it into .R format):

Project R Code Manual

R files of this project are given in below table.

Name	Description
IS580_FINALPROJECT.R	Main File
changefactorsToOrdered.R	Changes the attributes type ordered from factor
combineAndTrasformData.R	Combines the ordered and numeric attributes in one data frame
deleteOverMissData.R	Deletes the attributes with over 25000 missing value
doFactorImputationWithMICE.R	Imputation for factor attributes with MICE
doNormalization.R	Normalized the data with min-max, z-score and log methods
doNumericImputationWithAmelia.R	Imputation for numeric attributes with Amelia
performChi2Test.R	Is used to find some interesting attributes with over 25000 NAs
Model.R	SVM model

IS580_FINALPROJECT.R is the main file and must be run after the initialization of other files.

In order to test the model, normalizedSelectedSubset.csv file must be in working directory.