Forecasting Methods – Powerpoint Presentation

Introduction

Forecasting refers to refers to a strategy that attempts to predict the future using the past. It uses different data sources to arrive at a conclusion.
Forecasting provides entrepreneurs with vital information that helps them to run their businesses effectively. Most organizations use the results of their forecasting models in making various strategic decisions. Forecasting also helps financial experts to determine the financial position of an entity in future (Rao & Krishna, 2009).
There are several forecasting methods. They may range from simple manual spreadsheets to complex programs that integrate a wide variety of data to predict the future.

Characteristics of a good forecasting method

A good forecasting model should provide data that has a high degree of accuracy. Therefore, it should consider all factors that may affect the future in arriving at a conclusion.
A good forecasting model should provide information in a timely manner. The predictions of the model should relate with the future outcomes according to the time predicted by the model.
A good forecasting model should be easy to use and understand. It is a fact that some forecasting models may be complex. However, if this is the case, the forecasting method should be easy to use.

A good forecasting method should be reliable. The margin of error of the method should not be too large (Yaffee & McGeeAn, 2001)

Type of forecasting methods

Qualitative methods include forecasting techniques that use opinions, emotions, personal experiences in predicting the future. They are generally subjective. In addition, they do not rely on complex mathematical computations.
Qualitative methods refer to forecasting techniques that use mathematical models to predict the future. These techniques are usually subjective. In addition, they rely on complex mathematical computations.

Qualitative methods

Qualitative methods include executive opinion, market survey, sales force composite, and Delphi method.
Executive opinion refers to the situation where the management of develops a forecast after conducting a meeting.
Market survey is one of the most common forecasting methods that businesses use. It involves interviews to determine customer preferences. The businesses use the result of the survey to develop a forecast.
Sales force composite refers to a forecasting technique where the sales force develop sales estimates for their respective regions.
On the other hand, the Delphi method refers to a forecasting method where a group of experts form a consensus (Armstrong, 2001).

Quantitative methods

Quantitative forecasting methods include time series and causal models. Time series uses the patters of past data to determine forecast the future. On the other hand, associative methods try to predict the future by determining how other variables in the environment affect the variable under investigation.

Causal methods

Causal methods strive to determine the factors that influence the changes in various variables. The consider the seasonal variations in the data under investigation.
Most causal methods use strict algorithms to develop accurate forecasts. However, certain causal methods do not use the strict algorithms. Instead they rely on the forecaster’s judgment on the relationship between different variables. The author may use past relationships of variables to develop forecasts without having a full understanding of the interrelationship between the variables.
The most common causal methods include regression analysis and Autoregressive moving average with exogenous inputs (ARMAX)

Time Series Model

Time series is the most common quantitative forecasting method. Time series refers to a series of data points that are measured after even intervals. The Dow Jones Industrial Average is a good example of time series as it is measured after a uniform time – one day.
Time series is usually represented using a flow chart. The flow chart tracks the changes in he variable with time.
Forecasters strive to use time series models to identify patterns in certain variables. The variables may have a trend or occur in cycles. In addition, the variables may portray seasonality. In some instances the variables may portray randomness. This may it difficult to provide accurate forecasts (Montgomery, Jennings, & Kulahci, 2011).

Forecast of stationary time series

Stationary time series is slightly different from forecasting for conventional time series. Conventional time series have noticeable trends that may help forecasters in making future predictions. However, stationary time series do not have any noticeable trends. This makes it difficult to apply forecasting mode
Moving averages and exponential smoothing are the most common methods that are used in forecasting of stationary time series
ls of conventional time series to stationary time series.

Exponential smoothing

Exponential smoothing is a method that attempts to smooth a large amount of data making it easy to present the data. The data may be random or smooth but occur in a chaotic environment. This enables exponential smoothing to provide forecasts for data that would otherwise be impossible to predict.
Exponential smoothing assigns a set of exponentially reducing weights to previous data. However, the weights must add up to 1.
Below is the expression for exponential smoothing;

New forecast = Last period’s forecast + a(Last period’s actual demand – Last period’s forecast). Representing it in symbols, Ft = Ft-1 + a(At-1 – Ft-1) where a is the weight of the data (Hyndman & Athanasopoulos, 2014).

Get Your Custom Paper From Professional Writers. 100% Plagiarism Free, No AI Generated Content and Good Grade Guarantee. We Have Experts In All Subjects.

Place Your Order Now