Analysis of Machine Learning Technique to Predict Eggs Production in Poultry Farms
Loading...
Date
2022-09-01
Journal Title
Journal ISSN
Volume Title
Publisher
MUST Journal of Research and Development (MJRD)
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.
Description
This Journal Article was Published by Mbeya University of Science and Technology 2022