A hybrid genetic algorithm based approximate cash crop model with support vector machine classifier framework for predicting economic viability of underutilised crop in rural area

Oh, Mei Shin (2019) A hybrid genetic algorithm based approximate cash crop model with support vector machine classifier framework for predicting economic viability of underutilised crop in rural area. PhD thesis, University of Nottingham.

[img] PDF (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (2MB)

Abstract

The research deals with developing a hybrid Genetic Algorithm based on approximate cash crop model framework to predict economic viability of underutilised crop in rural area using Support Vector Machine. Machine learning techniques are dependent on large amount of data in order to predict outcomes. However, underutilised crops are, by definition, not cultivated on a large commercial scale causes the scarcity of training data for this application. In this proposed framework, SVM is implemented in conjunction with a Genetic Algorithm (GA) and associated fitness functions to generate training data for the SVM from approximate models developed for normal cash crops. Approximate models are used in Genetic Algorithm as fitness function to generate synthetic data. Synthetic data generated is used to train Support Vector Machine. Experiments are designed to compare synthetic data from actual models and data generated from approximate models. Model and real data from World Bank is also used to validate the proposed framework. Finally, synthetic data generated from approximate farm income models for crop are tested against the real village data obtained from Crops For the Future Research Centre. The result shows that good classification is attainable in spite of the inaccuracy that existed in the training data from an approximate model and artificially generated through Genetic Algorithm which includes constraints that reflect physical conditions found in rural villages. This framework is able to identify village which is able to achieve potential success economically planting underutilised crops before cultivating the crop itself based on the approximate model from village who has successfully commercialised normal crops.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Chen, ZhiYuan
Mohd Isa, Dino Isa Amshah
Subjects: Q Science > QA Mathematics > QA 75 Electronic computers. Computer science
Faculties/Schools: University of Nottingham, Malaysia > Faculty of Science and Engineering — Science > School of Computer Science
Item ID: 55811
Depositing User: OH, MEI SHIN
Date Deposited: 27 Feb 2019 04:40
Last Modified: 07 May 2020 13:01
URI: https://eprints.nottingham.ac.uk/id/eprint/55811

Actions (Archive Staff Only)

Edit View Edit View