Predicting Possition from Accelerometers Data through Stochastic Gradient Boosting
John Sprockel
October 2, 2020
For this project the prediction of the position will be carried out using accelerometer data developed in the practical machine learning course of Johns Hopkin also in coursera. This model is: Stochastic Gradient Boosting and was trained through the caret package in R. The variables that the model takes into account are:
- roll_belt
- accel_belt_z
- magnet_belt_y
- magnet_dumbbell_y
- roll_forearm
- pitch_forearm
Shiny Application
With this shiny prediction app, you can insert data from acelerometers and get back predictions about one of five body possitions. The model is based in a Stochastic Gradient Boosting.
The project info and ML development is in:
Link to the app
You can have access by the next link:
https://jjsprockel.shinyapps.io/Possition-Prediction-Shiny/
- Reference
Ugulino, W.; Cardador, D.; Vega, K.; Velloso, E.; Milidiu, R.; Fuks, H. Wearable Computing: Accelerometers’ Data Classification of Body Postures and Movements. Proceedings of 21st Brazilian Symposium on Artificial Intelligence. Advances in Artificial Intelligence - SBIA 2012. In: Lecture Notes in Computer Science. , pp. 52-61. Curitiba, PR: Springer Berlin / Heidelberg, 2012. ISBN 978-3-642-34458-9. DOI: 10.1007/978-3-642-34459-6_6.(http://groupware.les.inf.puc-rio.br/har)