Forecasting

Introduction To Forecasting

Forecasting and its need

Types of Forecasting

Steps involved in Forecasting

Types of Plots – Scatter Plot, Time Plot, Lag Plot, ACF Plot

Autocorrelation and Standard Error

Common pitfalls of plots and Aspect ratio

Time Series Components – Trend, Cyclical, Seasonal and Irregular

Ljung box test for identifying randomness

Forecasting Errors

Forecasting errors and the measures associated with it

Mean Error

Mean Absolute Deviation

Mean Squared Error

Root Mean Squared Error

Mean Percentage Error

Mean Absolute Percentage Error

Forecasting Methods

Forecasting methods based on smoothing

Moving Average

Exponential Smoothing

Decomposition of Time Series into 4 components

Additive Model

Multiplicative Model

Mixed Model

Curve fitting – Least square method

Simple exponential smoothing (SES)

Forecasting strategy – Separate, Forecast and Combine

Smoothing Methods

Moving Averages

Naive Model

Naive Trend Model

Simple Average Model

Moving Average over K time periods

Exponential smoothing

Simple exponential smoothing

Holt’s version

Winter’s modification

Modeling Different Components

Modeling Random Component

Models for Stationary Time Series

Autoregressive Model (AR)

Moving Average Model (MA)

Autoregressive Moving Average (ARMA) Model

Autoregressive Integrated Moving average (ARIMA) Model

Building seasonality into ARIMA models

Simple Linear, Multiple and Weighted Regression

Detecting Anomalies

Non-linearity detection

Scatter plot

Partial residual plot

Partial regression plot

Non-normality detection

Normal plot

Jarque-Bera Normality test

Transformations

Box-Cox

Box-Tidwell

Growth curve – Trend, Linear, Quadratic, Exponential and Sigmoid

ARCH & GARCH models