What machine learning techniques could be used to derive benchmarks for a time series data?

timeseries
benchmark_solution
machine_learning

#1

Objective is to derive benchmarks to optimize the existing alert system.


#2

Hi @SSingh

You can use models like ARIMA built for Time - series modelling or could try typical supervised ML techniques by creating lag features.

To start with its important to understand the basics of Time series. You can start with:

Best,
Saurav.


#3

Thanks for your reply @sauravkaushik8 .

How to use the output of forecasting to derive benchmarks for the data? Can we use delta (difference between predicted and actual value) to create lower and upper benchmarks or can anyone suggest some better alternative?