Time Series 🕐

Time Series forecasting is an application of Data Science

Every business operates under risk (based on market condition) and uncertainity (due to some natural calamity)

Forecasting helps us to assess these risks, and based on that the budgets can be managed.

In simple time we can understand with a following equation :

y = f(x)

y : Dependent Variable (Future)
x : Independent Variable (Past)

The data points should be at regular interval (The time interval should be regular)
Intervals of Time Series : Yearly, Quarterly (4), Monthly(12), Weekly(52), Daily(365), Hourly
Order matters in time series analysis.
No data should be missing in the given time intervals.

Time Series
Systematic Component		Irregular Component
Trend	Seasonality	Unpredicted calamity (Error)
Long term (Ups and downs)	Regular pattern	COVID pandemic or any uncertain natural calamity

Breaking of Time Series data into trend, seasonality and irregular components.

We have to find this three components in a given Time Series data.

There are two types of decomposition models.

Additive Model ( If Seasonality is constant ) : Observation = Trend + Seasonality + Error

Forecasting Sales with Trend, Seasonality and Error :

Sales = Trend (Business Growth) + Seasonality (Weather) + Error (Theft / Calamity)

Multiplicative Model ( If Seasonality is changing ) : Observation = Trend * Seasonality * Error

Python library that provides classes and functions to perform statistical tests on time series data.

iamkirankumaryadav/TimeSeries