Machine learning project. Six subjects performed several repetitions of an arm curl exercise. Each performance was executed either correctly or in a manner exhibiting one of four incorrect exercise actions. The subjects were instrumented with various inertial motion sensors and magnetometers. The time-series and aggregated data from the instruments were collected and labeled with the observed exercise action. A subset of the data was used to train various classifiers to predict exercise action on new data. The performance of the classifiers was estimated, the the best classifier was used to predict the exercise action on a leave-out set.

Predicts most likely immediate next word from a given phrase, based on a language model based on a statistical analysis of three source documents.

This document presents the current status of the text prediction project. Some basic statistics and observations about the source documents are included, as well as some analysis of N-gram tokenization, and thoughts about how to handle certain language constructs such as money, dates, and URLs. An outline of the prediction method and its implementation is presented, along with a list of the next steps to be taken for this project.

Brief documentation to the sales analysis tool at: https://markatango.shiny.io/project and also at my Shiny Server: http://data-dancer.com/project

This report presents data showing annual economic damage and adverse health effects attributed to recorded weather events from 1950 to 2011.