Comparative revision of artificial intelligent methods for discovering gene expression / Mohd Dzulkarnain Zainal Ahbiddin

Gene expression analysis is one of the studies in bioinformatics. There are
many methods and approaches use in gene expression analysis. Some methods that
are currently being used, such as fuzzy ART, neural network, and Bayesian method
were used in gene expression analysis. The problem that occurred in supervised
learning is that the output and error rates that been provided were momentous. The
reason in conducting this research is to determine the best methods, between two
approaches for gene expression analysis. For this research, the approach used is
supervised learning and the methods that were used are multi layer feedforward and
k-Nearest Neighbour. The methodology that will be use for this research are
knowledge acquisition, implementation that consists of three phase; experiment,
result and analysis, experiments and observation, and evaluation and findings. After a
series of experiments, the multi layer feedforward is the better method in determining
gene expression especially protein genes rather than k-Nearest Neighbour. It is
because the accuracy of the output is more precise and can be used for further
analysis. The presentation of multilayer feedforward is clear and well-defined. The
accuracy of the results is important for usage of others. This research can be a good
reference for the advancement and development of gene expression analysis.