Initial Draft - HimajaYerra/Masters-Dissertation GitHub Wiki
Audio data - it is the array of sound pressure amplitudes at a constant sample rate.
Sample rate - Time series data - It is the series of data points represented in a timely fashion.
Time series Analysis - These methods are used to study the behavior of time series data while, Time Series Forecasting methods are used to predict the future data points of a given time series.
In this project, we are using Time series analysis to in defining the characteristics of patient breadth data and mark each data point to either reliable or non- reliable.
Later, the classified data is used as the training data for a classifier that will predict the data reliability of data points in unseen data
For the model:
Features - Data Points sampled at appropriate frequency Label - Reliability measure
Time - series Anomaly detection
https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/46283.pdf
For literature review - Wavenet paper explained in detail https://arxiv.org/abs/1611.09482 Wavenet model implementation: https://github.com/usernaamee/keras-wavenet/blob/master/simple-generative-model.py