What is it about?

As susceptibility of commercial crops to the changing climates and resulting harsher conditions increases, interest in the potential of resilient underutilised crops grows. Therefore, alternative options are needed to be developed to mitigate the dangers associated with crop failure due to prevalence of disease and changing weather conditions that causes drought and loss of fertility in arable lands. Furthermore, It is also important to identify commercialisation potential of underutilised crops apart from biophysical and land qualities. The economic and demographic characteristics of an area which is conducive for normal commercial crops can be used to benchmark commercialisation potentials for underutilised crops. Hence, this project aims to assess the commercialisation possibility of underutilised crops on a large scale for under-developed areas with currently no possibility of growing commercial crops probably due to climate, soil characteristics. Support Vector Machine (SVM) method was implemented in conjunction with Genetic Algorithm (GA) and associated fitness functions to generate training data from approximate models which was developed for normal cash crops. The results showed that accurate classifications are obtainable even when the training is done with data from approximate and artificially generated data through implementation of a Genetic algorithm which includes constraints that reflect physical conditions found in rural villages. The simulation results showed that SVM is capable of acting as a filter for the inaccuracy in training data which is inherently present in the approximate models, thus allowing for better classification to be done on training data. This method can be used for rapid assessment of commercialisation potentials of underutilised crops in rural development programmes.

Featured Image

Why is it important?

As susceptibility of commercial crops to the changing climates and resulting harsher conditions increases, interest in the potential of resilient underutilised crops grows. Therefore, alternative options are needed to be developed to mitigate the dangers associated with crop failure due to prevalence of disease and changing weather conditions that causes drought and loss of fertility in arable lands. Furthermore, It is also important to identify commercialisation potential of underutilised crops apart from biophysical and land qualities.

Perspectives

The results showed that accurate classifications are obtainable even when the training is done with data from approximate and artificially generated data through implementation of a Genetic algorithm which includes constraints that reflect physical conditions found in rural villages. The simulation results showed that SVM is capable of acting as a filter for the inaccuracy in training data which is inherently present in the approximate models, thus allowing for better classification to be done on training data. This method can be used for rapid assessment of commercialisation potentials of underutilised crops in rural development programmes.

Dr zhiyuan chen
The University of Nottingham Malaysia

Read the Original

This page is a summary of: An economic feasibility assessment framework for underutilised crops using Support Vector Machine, Computers and Electronics in Agriculture, January 2020, Elsevier, DOI: 10.1016/j.compag.2019.105116.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page