Publications

Abstract : We gather enormous volumes of data to train a model, which aids a machine's ability to learn better. Not all of this data, though, will be applicable to the model. Classes or parts of the data can be removed if they don't significantly improve our model. The model may run slowly if there is too much extra data. It's also likely that the model will become erroneous as a result of learning from this unimportant data. By using only pertinent data and removing irrelevant data, feature selection reduces the input variable of the model. Using feature selection, we may improve our model in a number of ways, including preventing over-fitting and learning from noise, increasing accuracy, and cutting training time. Three types of feature selection methodologies―the filter technique, the wrapper approach, and the embedded method―have been developed throughout the years. In this study, we present the Moth Flame Optimization Method (MFOA) algorithm for gene selection from the Gene Expression dataset. A population-based algorithm with natural inspirations is called the Moth Flame Optimization Algorithm (MFOA). It takes its cues from how moths behave in the wild. The Moth Flame Optimization Algorithm (MFOA) can converge more Quickly and requires less compute than earlier methods.