Dr. Sahibzada Ali Mahmud

Enhancing Growth Curve Approach using CGPANN for Predicting the Sustainability of New Food Product

An enhancement to the GROWTH CURVE APPROACH based on neuro evolution is proposed to develop various forecasting models

To investigate the state and worth of the producer, to market a new product

Obtained using a newly introduced neuro evolutionary approach called Cartesian Genetic Programming based ANN (CGPANN)

An accurate and computationally efficient model is obtained, achieving an

accuracy as high as 93.37% on the time devised terrains, providing a general mechanism for forecasting models in mathematical agreement to its application in econometrics.

Comparison with other contemporary model evidences the perfection of the proposed model thus its vital power in developing the growth curve

approach for Predicting The Sustainability Of New Products.

Abstract

Marketing a product needs knowledge about the anticipation of the end user and forecasts make it possible to respond to these changes.

Different types of forecasts are made in order to mend the stability of a firm in the market including:

• Demand forecast

• Supply forecast

• Sales forecast

• Production forecast

• Profit forecast

• Advertisement and promotional expense forecast

Introduction

If the product that is to be commercialized is the RENOVATION of some already existing product(s) or product that were developed in the past, the deployment should follow the production and consumption trends of these previous versions of the products.

Techniques used for forecasting Trends of the market inclination includes;

• Stochastic techniques

• Probabilistic methods

• Machine learning applications

Situation

Cartesian Genetic Programming Evolved ANNCGP utilizes a two dimensional programming architecture that is incorporated by nodes or genes

CGPANN (Cont.)

In CGP, a genotype is represented by a string of integers with the corresponding phenotype a two dimensional nodal network.

The genotype is evolved by changing the connectivity and functions of nodes in the network, thus obtaining a range of topologies.

CGP evolved ANN

Input to a particular node of a genotype ‘g’

Output of a particular node ‘p’

• Nestlé

• Deans

• Saputo

Food Product Growth – Case Study

http://finance.yahoo.com/q?s=DF

https://ca.finance.yahoo.com/q?s=SAP.TO

http://finance.yahoo.com/q?s=NESN.VX

Simulation Attributes• Number of inputs taken (14-inputs/7days)

• Training Data Set (1 year)

• Accuracy and Error Calculation

• Fitness

Sliding Window Mechanism

Training Accuracy

MAPE For training session for predicting single, half week and full week future instances, using historical data of 1 week

Network initially with 450 instances of weekly stock worth as its input exhibits 1.79% Mean Absolute %age Error.

Testing Results

MAPE results for different data sets - 450 initial nodes - 14 instances input- Single output

Network initialized with 450 nodes and 14 instances Market values as its input has 11% MAE that descends to 5.3% of MAPE for different stock market entities.

Comparison with Other Models

Comparison of proposed CGPANN model with different models

CGP based ANN has its lead over other state of the are models for its 93.37% Mean Absolute Accuracy, depicting its perfection and adoptability

Actual Verses Forecasted Terrain

Predicted growth curve using CGPANN (initialized with 450 nodes)

• The paper uses CGPANN prediction model for forecasting the growth curve of a firm

• The buying power of the consumer can be predicted by looking at the daily change in the stock usage that can be used in the estimation of the future production by a company

• The producers can then manage their inventory while knowing the exact number of future consumer of each and every single product, supplied by them.

• With a model having 93.37% accuracy, one can go for a finest decision while launching a new product or continue marketing previous.

Conclusion

Questions!!