Abstract
Mushrooms are one of the groups of living organisms in the fungal regnum which have umbrella-like body characteristics. The body consists of an upright part that functions as a rod to support the hood as well as a hood that is horizontal and rounded with different color variations. There are types of mushrooms that can be a food source for humans. Some types of mushrooms can be eaten or processed like other foods. Apart from that, some types of mushrooms are dangerous if consumed by humans because they are poisonous. Based on these problems, this study offers a new contribution in identifying types of poisonous and non-toxic mushrooms based on mushroom characteristics using the CART algorithm combined with the LogitBoost boosting algorithm. The aim of this research can be used as material for further studies in making tools that can effectively and accurately differentiate between poisonous and non-toxic types of mushrooms. This can help reduce cases of poisoning due to consumption of poisonous mushrooms. The data used is secondary data from public sources UCI Machine Learning Repository. Evaluation of model performance resulted in an accuracy of 98.79%; recall 98.70%; specificity 98.85%; precision 98.56%; F1-Score 98.63%, and AUC 0.9876. These results show that the model is very effective in detecting poisonous mushrooms and has minimal errors in classification.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call