/ZLG-ML_Project_Spring2018

TalkingData AdTracking Fraud Detection

Primary LanguagePythonMIT LicenseMIT

ZLG-ML_Project_Spring2018

Brandeis_CS123a_FinalProject In kaggledata.mat, there are the feautes of click_time (year,month,day). The time of day is in string array called "time".

trainsample is the original data and it is a table. X is the features of trainsample and is a table y is the result of trainsample and is an array of doubles XNew contains the columns (in this order) ip,app,device,os,channel,datenum (of the dates) and is an array of doubles

Work Distribution:

Brian Gao data preprocessing & model testing
Ruoshi Liu data augmentation & post processing
Han Zheng model selection & parameter tuning