/time-series-exercises

Time Series Machine Learning Exercises

Primary LanguageJupyter Notebook

Time Series Exercises

Summary

  • Time series analysis is about finding patterns in temporal data and making predictions, forecasting, based on those patterns.
  • What should we expect to happen next? (Time Series Analysis)

Common Use Cases

  • Sales & Revenue
  • Signal Processing
  • Speech & Chatbots
  • Economics
  • Healthcare
  • Stock Values
  • Census Analysis
  • Supply Chain
  • Staffing

Time series vocabulary

Vocabulary Definition
Temporal Relating to time
Periodic Occurring at intervals
Resampling in Time Sereis Changing the frequency of your data points
Stationary Process Distribution does not change over time
Trend Long term progression (increasing, decreasing, e.g.)
Seasonality Changes in patterns due to seasonal factors
Heteroskedasticity Changes in variance over time
Autocorrelation 'Regression of self', used to detect non-randomness in data. It is a correlation coefficient, but instead of between two different variables, it is between the values of the same variable at two different times
Lag Variables Previous time steps