Demand Forecasting-2(A Division)

Views:
 
Category: Education
     
 

Presentation Description

No description available.

Comments

By: gguerrer (114 month(s) ago)

Would you send this ppt to my mail pls? [email protected]

By: esingh91 (117 month(s) ago)

plz send this ppt to my mail id: [email protected]

By: vikkyshenoy (118 month(s) ago)

pls send this file to my e-mail ID vikkyshenoy @gmail.com

By: bibzvilavinal (122 month(s) ago)

pls send this ppt to my mail id,, [email protected]

By: bluemecha (130 month(s) ago)

mail these ppt to [email protected]

See all

Presentation Transcript

Slide 1: 

Demand Forecasting-2

Demand forecasting of established products : 

Demand forecasting of established products Mechanical extrapolation techniques or time series analysis Barometric (Or, Leading Indicator Technique) Statistical Methods Econometric Models Simultaneous Equation Method

Demand Forecasting of new Products : 

Demand Forecasting of new Products Survey of buyers’ intensions Test marketing Life cycle segmentation analysis Bounding curves method

Mechanical extrapolation techniqueor Time series analysis : 

Mechanical extrapolation techniqueor Time series analysis Meaning :- Whenever data are collected and arranged in accordance with time, in the form of series is known as time series or trend projection method.

Slide 5: 

Assumption :- It is based on the assumption that future events are a continuation of the past and, therefore, historical data can be used to predict the future.

Utility of time series analysis : 

Utility of time series analysis It helps to understand the past behavior and present situation and also to predict the future values i.e., to find trend value for a specific year. To find seasonal fluctuation in the variable. To businessmen and industrialists the study of the time series of demand, supply, price, production, sales etc is important to make the necessary adjustments in their business policies.

Components of time series : 

Components of time series Variations in time series long term short term variations variations Trend cyclical seasonal irregular

Slide 8: 

Hence the total variation = Trend + Seasonal variations + Cyclical variations + Irregular variation i.e., O = T+S+C+I

Slide 9: 

Trend variation :- If a time series is observed, there may appear many ups and downs in the series. However if it is carefully examined, a persistent growth or decline may be found there. This tendency is known as trend. Cyclical component : It is a wavelike, repetitive moment fluctuating about trend of the series. e.g., depression, recovery, prosperity and recession in general business activity repeat themselves after some time. The minimum time period of a cycle is greater than one year.

Slide 10: 

Seasonal variation :- Fluctuations that always occur at a particular time of the year. e.g., sales of woolen clothes are higher during winter season. Sales of umbrella will be highest during monsoon. Irregular variation :- These are those variations which are left over after isolating the other components. All type of variations other than trend, seasonal and cyclical variations can be termed as irregular variation. E.g., strikes, floods, natural calamities, war etc

Decomposition of time series : 

Decomposition of time series The four components may be related to each other in an additive or multiplicative form :- O = T+S+C+I (Additive model) O= T * S * C * I (Multiplicative model) We can convert multiplicative model into additive model by taking logarithms as: log O = log T + log S + log C + log I

Graphical Method : 

Graphical Method It is also known as Fitting Trend Line by Observation Graph paper: (1) Arithmetic scale (2) Semi-logarithm scale

Slide 13: 

Advantage Easy & Quick method Disadvantage Lacks of Scientific Temper

Least-Square Method : 

Least-Square Method Statistical Method Forecast Sales Independent Variable: '' Time'' ''naive''- because doesn't explains the reasons of change. Least-Square Method:(1) Linear Trend (2) Non- linear Trend

Linear Trend Method : 

Linear Trend Method Also known as Straight-line Trend S = a + b. T where, a & b = constant, T = Trend ƸS = Na + b Ƹ T (N = no. of. years) ƸST = a ƸT + b Ƹ T2

Non-Linear Method : 

Non-Linear Method (a) Second & higher degree polynomial trends S = a + bT+cT2 (b) Third degree/ Cubic Trend S = a + bT+cT2 + dT3 Exponential Trend/ Semi-log trend S = a ebt log S = log a + log bT Iv Double-log Trend S = aT b log S = log a + b logT

Slide 17: 

Smoothing Methods

Slide 18: 

In cases in which the time series is fairly stable and has no significant trend, seasonal, or cyclical effects, one can use Smoothing methods To average out the irregular components of the time series.

Slide 19: 

Moving Average Method

Slide 20: 

Moving Average Method The moving average method consists of computing an average of the most recent n data values for the series and using this average for forecasting the value of the time series for the next period..

Simple Moving Average : 

Simple Moving Average Include n most recent observations Weight equally Ignore older observations weight today 1 2 3 ... n 1/n

Slide 23: 

Issues with moving average forecasts: All n past observations treated equally; Observations older than n are not included at all; Requires that n past observations be retained; Problem when 1000's of items are being forecast.

Slide 24: 

Exponential Smoothing

Slide 25: 

Include all past observations Weight recent observations much more heavily than very old observations

Exponential Smoothing : 

Exponential Smoothing weight today Decreasing weight given to older observations

Exponential Smoothing : 

Exponential Smoothing weight today Decreasing weight given to older observations

Exponential Smoothing : 

Exponential Smoothing weight today Decreasing weight given to older observations

Exponential Smoothing : 

Exponential Smoothing weight today Decreasing weight given to older observations

Exponential Smoothing : 

Exponential Smoothing weight today Decreasing weight given to older observations

Slide 31: 

Ft+1= α Xt + (1- α) Ft α is the smoothing constant (a number between 0 and 1) Ft is the forecast for period t Ft +1 is the forecast for period t+1 Xt is the actual data value for period t

Slide 33: 

For Exponential Smoothing, how do I determine the best value of α

Slide 34: 

Find Out Ft Then Find Out Ft-Xt Find Square of Ft-Xt Make Total Of above Less Total=Preferable Weight

Slide 35: 

ARIMA Method Box-Jenkins methodology

Slide 36: 

Auto-Regressive Integreted Moving Averages (Smoothing Methods + Auto-Regressive Methods)

Slide 37: 

Used & Useful When Short-Term Forecasting is to be made Inherrent Patern of Trend is Highly Complex & Difficult

Slide 38: 

5 Step Process Removal Of Trend Parameter Estimation Model Identification Verification Forecasting

Slide 39: 

Removal Of Trend Useful for Stationary Time Series Remove Trend Component by Differencing Methods

Slide 40: 

Model Identification Find Combinations from below 3 1.Order of Involvment of Autoregressive terms 2.No.of Differences to remove orignal trend 3.Use Moving Average Terms Whichever fits best to the Time Series

Slide 41: 

Parameter Estimation Find a Particular Combinations Use Method of Least Square Coefficients are thus obtained

Slide 42: 

Verification Check the Goodness of the Fit of Model Do Residuals show specific Pattern? -Yes - Go Back to Step2 -No - “Good” Fit

Slide 43: 

Forecasting Arrive at Good fit Use Coefficient to Forecast Use of Softwares for forecasting

Slide 44: 

Limitations Of Softwares Very Expensive Requires Experts & Qualified Forecasters Difficulty in Operation of this Model

BEROMETRIC TECHNIQUE : 

BEROMETRIC TECHNIQUE

WHAT IS BAROMETER? : 

WHAT IS BAROMETER?

Barometer is an instrument used to measure the atmospheric pressure.Like that this method is used to forecast the demand of established product. : 

Barometer is an instrument used to measure the atmospheric pressure.Like that this method is used to forecast the demand of established product.

ASSUMPTIONS : 

ASSUMPTIONS Relationship can exist among various economic time series. For e.g. Industrial production overtime and industrial loans by commercial banks overtime. 2. Future can be predicted from certain events occurring in the present. For e.g. Birth rate at present & demand for seat in school

UEFULNESS : 

UEFULNESS PROVIES AN INDICATION FOR DIRECTION OF CHANGE

3 Indicators : 

3 Indicators There are 3 indicators in barometric technique which act as a complement to each other. i.e. 1. Leading indicator 2. Coincident indicator 3. Lagging indicator

Leading Indicator : 

Leading Indicator Variables whose movement precede the movement of some other variable For e.g. Series of Birth rate & series for enrolment in schools Application for amt of housing loan & demand for construction material

Coincident Indicator : 

Coincident Indicator When data of one variable moves along with data of another variable for e.g. Series of data on national income & series of data on employment in an economy

Lagging Indicators : 

Lagging Indicators Where data moves behind the series being compared. For e.g. Series of number of babies , lagging series for the series of marriage

Forecasting stages : 

Forecasting stages To locate the leading indicator for the variable to be forecast To estimate a mathematical relationship To find out the forecasted values of the variable and If possible, to verify the validity of the forecast

For e.g. : 

For e.g. X is a number of children born on particular day in a town, and Y is the demand for seats in schools (where the required age at the time of admission is exactly 5 years) in a town.

EquationYt = f (Xt -s)where, t refers to current years ands refers to length of the lag : 

EquationYt = f (Xt -s)where, t refers to current years ands refers to length of the lag

Limitations of leading indicator : 

Limitations of leading indicator Not easy to identify a proper leading indicator It tells us about the direction of change and not about the magnitude of change.

Index numbers : 

Index numbers Diffusion indices (group of indicators) Composite indicators (weighted avg)

Slide 59: 

STATISTICAL METHODS OF FORCASTING

Slide 60: 

(1) NAÏVE MODELS (2)CORRELATION AND REGRESSION METHOD

Slide 61: 

3-61 Naive Forecasts The forecast for any period equals the previous period’s actual value.

Slide 62: 

Quantitative forecasting methods in library management 62 Naive forecasts NF1 and NF2 Naive forecasts: (a “folk forecasting technique”) NF1 means “The value tomorow will be the same as today”. Example: Number of library visitors today was 120. Forecast NF1 for tomorow: 120. NF2 means “The value tomorow will be less /greater by …10% ”. Example: Average temperature this month is 20 degrees. Forecast NF2 for the next month: Temperature will be 25 degrees (increase of 25%).

Slide 63: 

Simple to use Virtually no cost Quick and easy to prepare Easily understandable Can be a standard for accuracy Cannot provide high accuracy Naive Forecasts

Slide 64: 

Stable time series data F(t) = A(t-1) Seasonal variations F(t) = A(t-n) Data with trends F(t) = A(t-1) + (A(t-1) – A(t-2)) Uses for Naive Forecasts

Slide 65: 

MOST POPULAR METHODS UNLIKE TIME SERIES ANALYSIS CORRELATION AND REGRATION ANALYSIS DOSE NOT LIMIT IT SELF TO “TIME” AS THE INDEPENDENT VARIABLE CORRELATION AND REGRESSION METHODS

Slide 66: 

MEANING- THE RELATIONSHIP BETWEEN TWO VARIBLES SUCH THAT A CHANGE IN ONE VARIBLE RESULT IN A POSITIVE OR NEGATIVE CHANGE IN THE OTHER VARIABLE OR DIRECT OR INDIRECT CAUSE AND EFFECT RELATIONSHIP BETWEEN TWO AND MORE VARIABLES IS CALLED CORRELATION CORRELATION ANALYSIS

Slide 67: 

POSITIVE CORRELATION— TWO SERIES VARYING IN THE SAME DIRECTION (r>0) NEGATIVE CORRELATION— TWO SERIES VARYING IN THE OPPOSITE DIRECTION (r<0)`

TWO TYPES OF CORRELATION ANALYSIS : 

TWO TYPES OF CORRELATION ANALYSIS SIMPLE OR PARTIAL CORRELATION WHEN ONLY ONE INDEPENDENT VARIABLE IS TAKEN TO EXPLAIN VARITIONS IN DEPENDENT VARIABLE E.G-SALES & PRICE MULTIPLE CORRELATION WHEN THERE ARE MORE THAN ONE INDEPENDENT VARIABLE IS TAKEN TO EXPLAIN VARITIONS IN DEPENDENT VARIABLE E.G-SALES & PRICE,INCOME ADVERTISEMENT

COEFFICIENT OF CORRELATION : 

COEFFICIENT OF CORRELATION IT IS USED TO KNOW CLOSENESS OF VARIABLES

Slide 70: 

REGRESSION EQUTION METHODS (1) FITTING SIMPLE LINEAR REGRESSION (a) GRAPHICAL METHOD (b)LEAST SQUARES METHOD (2) FITTING SIMPLE NON-LINEAR REGRESSION EQUTION (a)LOGARITHMIC MODEL (b)PARABOLIC REGRESSION MODEL (C) MULTIPLE REGRESSION ANALYSIS

(1) FITTING SIMPLE LINEAR REGRESSION : 

(1) FITTING SIMPLE LINEAR REGRESSION (a) GRAPHICAL METHOD

Slide 73: 

(b)LEAST SQUARES METHOD S=a+Bp NORMAL EQUATION ∑S=na+b∑p ∑sp=a∑p+b∑p2 bsp=n∑SP- ∑S ∑P n∑P2-(∑p)2 a= ∑S-b ∑P n

(2) FITTING SIMPLE NON-LINEAR REGRESSION EQUTION : 

(2) FITTING SIMPLE NON-LINEAR REGRESSION EQUTION (a)LOGARITHMIC MODEL log S=a+b*p

(B)PARABOLIC REGRESSION MODEL : 

(B)PARABOLIC REGRESSION MODEL

(C) MULTIPLE REGRESSION ANALYSIS : 

(C) MULTIPLE REGRESSION ANALYSIS SALES= a, PRICE +b, ADVERTISING +c, RIVALS’ PRICE LEVEL +d, PERSONAL DISPOSABLE INCOME + u WHERE a,b,c,d are partial regression coefficient

Slide 77: 

LIMITATION (i) BASED ON PAST DATA (ii)ACCURACY OF MEASUREMENT OF INDEPENDENT VARIABLE DETERMINES THE DEGREE TO WHICH THE CONFIDENCE CAN BE PLACED IN THE FORECASTED VALUES OF THE DEPENDENT VARIABLE

ECONOMATRIC MODEL : 

ECONOMATRIC MODEL The main difference between time series and econometric model is- Time series- only time was taken as an independent variable Econometric model- all ecomonic and demographic variables which influece forecasting variable are taken as independent variable.

Example Qd=w + B1*P1T + B2*P2T + B3*A1T + B4*A2T + B5*WT + B6*YT + e Q=q dd of air condition of a given brand P1=price of given brand P2=price of sustitute A1=adv exp of a given brand A2=adv exp of rival brand W=weather condition Y=aggregate personal disposable income e=disturbance term

3 stages of econometric model : 

3 stages of econometric model Identification of variables and the functional form Estimation of parameters Finding the forecast values

Classification of econometric models : 

Classification of econometric models 1. Single equation models: Only one equation is used 2. Simultaneous models: Set of equations are used a. Endogenous variables values are determined by the model b. Exogenous variables values originate from outside the system

SIMULTANEOUS EQUATION MODEL : 

SIMULTANEOUS EQUATION MODEL When variables are dependent and independent at the same time, this model is used. Set of equations are used where a variable affects another variable and in turn is also affected by the other variable. This method is more complex. It is also known as COMPLETE SYSTEM APPROACH.

Example : 

Example Simultaneous model is consist of 2 type of equation- 1.Identities: represent relationships that are necessarily true 2.Behavioural equations: represent the expected behaviour of physical process or individuals E.g.: Yt = Ct + Gt + It Ct = b0 + b1*Yt + et It = a0 - a1*it + a2*Pt-1 + et

Advantages : 

Advantages Able to identify reason for the change Possible to forecast the effect of policy variables by tracing them as an independent variables Predicted values are compared with the actual data, it is possible to improve the accuracy of prediction of the model We need to estimate the future values of only predetermined variables (exogenous)

DEMAMND FORECASTING FOR NEW PRODUCTS : 

DEMAMND FORECASTING FOR NEW PRODUCTS Survey of buyer’s intention (or,consumer survey) Test marketing Life cycle segmentation analysis Bounding curves method

3. Life Cycle Segmentation : 

3. Life Cycle Segmentation

Stages in product life cycle : 

Stages in product life cycle Introduction: Quality, advertising, price, service Growth: Advertising, quality Maturity: Price, advertising, quality, service Saturation: product differentiation in quality, packaging, advertising Decline: Find new product-uses and advetise them, quality and service, price

4. Bounding curves method : 

4. Bounding curves method

authorStream Live Help