Date Log
Analysis of Machine Learning Technique to Predict Eggs Production in Poultry Farms
MUST JOURNAL OF RESEARCH AND DEVELOPMENT,
Vol. 2 No. 1 (2022): Analysis of Machine Learning Technique to Predict Eggs Production in Poultry Farms
Abstract
Machine learning techniques have emerged as a great tool for improving agriculture's economic activities. Machine Learning has specifically been applied in chicken farming to forecast egg output, enhancing both the economies of the farmers and the nation. In this research, we present a study devoted to the analysis of machine learning approaches to forecast egg output. The study looked at four (4) machine learning algorithms, the quantity of features utilized as input, and the strengths and weaknesses of each method. A number of features with a mean value of more than 6 are employed by an Artificial Neural Network, yet this network is unable to extract features from the dataset. Fuzzy logic uses many features with a mean value of more than 4.5 but few datasets. Few datasets and features with a mean value of less than 4 are used by Random Forest and Support Vector Machine. Compared to other techniques, Artificial Neural Network is the most popular and has a high mean value of features, but it is unable to extract core features from the dataset. Additionally, it only employs small datasets. The model's stability is reduced when limited features and datasets are used. Deep learning is built on the Artificial Neural Network, but so far only feedforward and backward architecture have been applied. It is obvious that poultry farmers would benefit from using machine learning to manage both their marketing and production processes. This study recommends the use of deep learning techniques with the best architecture due to the drawbacks of the currently existing techniques. These techniques will be able to employ numerous features and a large number of datasets, improving the stability of the model.
Keywords
Download Citation
Endnote/Zotero/Mendeley (RIS)BibTeX