#Practical Machine Learning: Project Writeup
The R script (classifier.R) does the following thing to solve this classification problem:
1.It loads the data from the given train and test sets.
2.It then does some data preprocessing like removing the first 6 columns. They contain dates, times, serial numbers etc. and hence we can get rid of them.
3.It then does some further pre-processing of the data by removing the columns which have NA's as these create problems while applying Machine Learning Algorithms.
4.After that it builds a random forest classifier with 500 trees using all the remaining columns.
5.The forest generated has an Out of Bag error of 0.13% and confusion matrix and OOB error vs number of trees as shown in the screenshot and plot respectively included in the repository.
6.The trained random forest model is then used to make predictions for the 20 test data points (All of which got accepted :) )
