Unsupervised Feature Selection for Time-Series Data

 By datasciencecentral - 2021-03-20

Description

Hello, friends. In this blog post, I will take you through an use case application scenario of the algorithms with my package “msda” for the time-series senso…

Summary

  • Hello, friends.
  • Determining the number of components An important part of using PCA is to estimate how many components are needed to describe the data.
  • The most appropriate sensors/features to be selected based on my variation-trend-capture-relationship approach would be then‘net_in’, ‘mem_util_percent’in the order of highest importance The reasons are as follows:- 1) It has a moderate number of values above the threshold value (i.e., in our case mean).
  • 2) The column values mostly remain constant or increase over time as seen from the slope.

 

Topics

  1. Backend (0.29)
  2. Database (0.16)
  3. Machine_Learning (0.12)

Similar Articles

Data Transformation: Standardization vs Normalization

By KDnuggets - 2021-03-14

Increasing accuracy in your models is often obtained through the first steps of data transformations. This guide explains the difference between the key feature scaling methods of standardization and ...

LSTM for time series prediction

By KDnuggets - 2021-01-23

Learn how to develop a LSTM neural network with PyTorch on trading data to predict future prices by mimicking actual values of the time series data.