Data in The Project

Data Source

In this mode, A Machine Learning model (Sensor Model) will give an evaluated score of the task using a file (contained the data/reading from seneors all over the house). For demo purpose, the model used a file on the server to make prediction. Our model use this data to train and test the model performance. In the evaluate phase, the model achieved 60% accuracy. The prediction is not an medical diagnosis, it just remind user to pay attention to health status.

Process Data

Before training the model, the original data was going through some essential processes in order to achieve better result. These processes included re-structure, feature selection and evulation. After the processes, different data files were categorized into two groups for its purpose. The origial data from sensor records looks like below figure.