My application developed for the Data Science Capstone (April, 2015)

In this report we present the initial study of the data involved in the capstone project of the Data Science Specialization focused in Text Mining and Natural Language Processing. For this we proceed as follows: Loading the datasets and sampling them for training our model. This leads to the cleaning procedure (two steps) and the tokenization process allowing us to identify in our sample 1,975 word types and a total word count of 774,005. Next we develop a Bigram Language Model using the tools of the R package “tm” and the word types found in the tokenization to finally identify 139,112 bigrams. This model with let us begin the predictive process and to evaluate the trade-offs between accuracy and memory consumption.

Peer Assessment 2 for the course Reproducible Research provided by The Johns Hopkins University through Coursera.

Project for the Developing Data Products Course

My first paper for the course Reproducible Research offered by Johns Hopkins University.