group15 wilkin case
Wilkin Case
Submitted By :
•Atrayee Bhattacharya FT151035
•Dhulipala Bharadwaj FT151008
•Tanmoy Bose FT151019
•Souvik Dey FT153079
•Soumendu Mukhopadhyay FT151034
•Vivek Anand FT153113
•Anand M FT152020
•Manzoor FT152099
•Lokesh Chandana FT152087
Case – Problems & Issues
Problems :
•To identify a forecasting technique so as to forecast the demand for existing products as well as new products.
Issues :
•Best fit model ??
•Problems associated with each forecasting model ??
•Performance of the used model ??
•Fundamental things to be accounted while looking at forecast performance ??
•Managerial issues to be addressed ??
•Quantification of data to capture the issues ??
Methods Available
•Static Methods : demand observation shows trend and seasonality •Dynamic methods:
Moving average: demand observation shows neither trend nor seasonality
Simple exponential : demand observation shows neither trend nor seasonality
Holt’s model : demand observation shows trend only
Winter’s model : demand observation shows trend and seasonality Multiple Linear Regression
Demand = Systematic + Random
Thus, Static method cannot be used as in this even if new demand is observed, the estimates of level, trend and seasonality within a systematic component do not vary.
•We have considered PVC family to do our analysis and
Multiple Linear Regression
•Multiple Linear Regression can also be carried upon the data
Results of regression Analysis
Regression between Quarterly sales and Unemployment
Rate, Bank Prime Loan Rate & Housing
R2 = 0.129
Adjusted R2 = -0.0887
P-value = 0.6316
At 95% confidence interval, the regression is not statistically significant.
The
values of R2 tells us that the percentage of fluctuation in the dependent variable, Quarterly sales explained by the predictor variables - Unemployment
Rate, Bank Prime Loan Rate and Housing starts is low hence we discard causal forecasting for predicting demand of fire valves.
Winter’s Model
•To take into account both trend &
Submitted By :
•Atrayee Bhattacharya FT151035
•Dhulipala Bharadwaj FT151008
•Tanmoy Bose FT151019
•Souvik Dey FT153079
•Soumendu Mukhopadhyay FT151034
•Vivek Anand FT153113
•Anand M FT152020
•Manzoor FT152099
•Lokesh Chandana FT152087
Case – Problems & Issues
Problems :
•To identify a forecasting technique so as to forecast the demand for existing products as well as new products.
Issues :
•Best fit model ??
•Problems associated with each forecasting model ??
•Performance of the used model ??
•Fundamental things to be accounted while looking at forecast performance ??
•Managerial issues to be addressed ??
•Quantification of data to capture the issues ??
Methods Available
•Static Methods : demand observation shows trend and seasonality •Dynamic methods:
Moving average: demand observation shows neither trend nor seasonality
Simple exponential : demand observation shows neither trend nor seasonality
Holt’s model : demand observation shows trend only
Winter’s model : demand observation shows trend and seasonality Multiple Linear Regression
Demand = Systematic + Random
Thus, Static method cannot be used as in this even if new demand is observed, the estimates of level, trend and seasonality within a systematic component do not vary.
•We have considered PVC family to do our analysis and
Multiple Linear Regression
•Multiple Linear Regression can also be carried upon the data
Results of regression Analysis
Regression between Quarterly sales and Unemployment
Rate, Bank Prime Loan Rate & Housing
R2 = 0.129
Adjusted R2 = -0.0887
P-value = 0.6316
At 95% confidence interval, the regression is not statistically significant.
The
values of R2 tells us that the percentage of fluctuation in the dependent variable, Quarterly sales explained by the predictor variables - Unemployment
Rate, Bank Prime Loan Rate and Housing starts is low hence we discard causal forecasting for predicting demand of fire valves.
Winter’s Model
•To take into account both trend &