Mushroom Recognition Using Neural Network

2022 11th MEDITERRANEAN CONFERENCE ON EMBEDDED COMPUTING, 2022

Of the millions of mushroom species growing all around the world, one type is edible, while the other is poisonous. It is not easy to distinguish edible and poisonous mushrooms from each other and it is a condition that requires expertise. The classification of poisonous and edible mushrooms is therefore important. Machine learning algorithms are an alternative method for classifying poisonous and edible mushrooms using morphological or physical features of fungi. The dataset used in this study is the Mushroom dataset available in the UC Irvine Machine Learning Repository. Based on 22 features in the Mushroom dataset and four different machine learning algorithms, models have been created for the classification of edible and poisonous fungi. The classification success rates of these models were obtained from Naive Bayes, Decision Tree, Support Vector Machine and AdaBoost algorithms with 90.99%, 98.82%, 99.98% and 100%, respectively. When these results were examined, taking into account the physical appearance features of the mushrooms, it was determined whether the mushrooms were edible and poisonous by 100% with the AdaBoost model.

Journal of Intelligent Software Systems

This study applied five types of Mushrooms, they are Button mushrooms, Wood Ear mushrooms, Straw mushrooms, Reishi mushrooms and Red Oyster mushrooms. The feature extraction used is Order 1 with the parameters of mean, skewness, variance, kurtosis, and entropy. The process carried out to identify mushroom images by preparing image objects. There were 15 images of each mushroom class were taken for each mushroom and stored in .jpg format. The image processing is carried out by a feature extraction process. Then five images for each mushroom class are chosen. They were used as test images which will be classified so that identification results are obtained. This study applies the Classification Algorithm C4.5 to build a decision tree, which will also identify the results of the accuracy of processed mushroom images. The obtained result of accuracy was 84% in the classification of feature extraction Order 1

Caspian Journal of Health Research

Background: Mushroom is one of the sources for protein supply, and it has taken into consideration among most countries in the world due to its rich medicinal features. Nowadays, due to the mechanization of traditional methods and quality control of products, it is possible to evaluate the quality of mushrooms with the help of image processing techniques. Methods: In this study, image processing systems were used to determine the appearance quality of mushrooms. Using the properties of color, area, weight, and volume obtained from data mining techniques, artificial neural networks and fuzzy logic system mushroom quality was evaluated. Results: A total of 250 images in three categories of defective, moderate were assessed. The correct detection rate by the image processing system was 95.6%. Conclusion: The results of this study showed the optimum performance of image processing systems for assessing the quality of mushrooms. The superiority of image processing systems compared to traditional method can be observed in the quality of increased efficiency and high accuracy, as well as the reduction of costs and destructive effects in the production and packaging of food products.

2018 International Conference on Information and Communications Technology (ICOIACT)

Indonesia has 13% species of mushroom in the world but there is a very limited study on determining edible or poisonous mushroom. Classification process of poisonous mushroom or not will be easily conducted by learning machine using mining data as one of the ways to extract computer assisted knowledge. Currently, there are three comparisons of the best classification algorithms in data mining, namely: Decision Tree (C4.5), NaïveBayes and Support Vector Machine (SVM). The study method used is experiment with assisted tool of WEKA that has been testing in the comparison of the three algorithms. To conduct the testing, it is used the mushroom data of Agaricus and Lepiota family. The mushroom data were taken from The Audubon Society Field Guide to North American Mushrooms, in UCI machine learning repository. Results of the testing indicate that the C4.5 algorithm has the same accuracy level to the SVM by 100% however, from the speed aspect, process of the C4.5 algorithm is faster than the SVM.

Advances in Intelligent Systems and Computing, 2013

The research offers a quite simple view of methods to classify edible and poisonous mushrooms. In fact, we are looking for not only classification methods but also for an application which supports experts' decisions. To achieve our aim, we will study different structures of neural nets and learning algorithms, and select the best one, according to the test results.

Journal of Applied Informatics and Computing

The purpose of this research is to classify the mushroom based on its characteristic to be in an edible class or poisonous one using the Decision Tree Algorithm. The result showed that odor is the most important attribute to classify the mushroom. Mushrooms which have almond and anise odors are edible, while the rest of it, such as pungent, foul, creosote, fishy, spicy, and musty are poisonous which means they can't be eaten. For mushrooms that have no odor, there are some attributes to be checked such as spore-print-color, gill-size, gill-spacing, and population. At first, overfitting happened. To overcome this, the researcher used Random Sampling Techniques until got better accuracy. The most accurate sample is 99,9% using sample 6 or 2000 data.

Mushrooms belong to fungi kingdom and they contain essential nutrients such as proteins, vitamins, minerals, amino acids, antibiotics and antioxidants. There are numerous health benefits of mushrooms. All species of mushrooms are not edible. So before consuming, it should be checked for edibility. Accurate determination and proper identification of species are the only safe way to ensure edibility, and safeguard against possible accident of consuming poisonous one. This paper aims in classifying edible and poisonous mushrooms based on certain attributes like shape, size, colour, etc. A large volume of data is downloaded from University of California, Irvine (UCI). repository. The data have been formed into two groups, one for training and another for testing. We use one of the Supervised learning Algorithms for training. This classification model is saved and then used later to predict whether a given mushroom is edible or not.

Nepal Journal of Mathematical Sciences

Mushroom is a popular fruit of a much larger fungus that has a high level of protein and a rich source of vitamin B. It aids in the prevention of cancer, weight loss, and immune system enhancement. There are numerous thousands of mushroom species within the world and a few are eatable and a few are noxious due to noteworthy poisons on them. Hence, it is a vital errand to distinguish between eatable and harmful mushrooms. This paper focuses on comparing the performance of Random Forest and Reduced Error Pruning (REP) Tree classification algorithms for the classification of edible and poisonous mushrooms. In this paper, mushroom dataset from UCI machine learning repository has been classified using Random Forest and REP Tree classifiers. The result based on accuracy, precision, recall and F-measure showed that the Random Forest outperformed REP Tree algorithm as it had highest accuracy value of 100%, precision value of 100%, recall value of 100% and F- measure value of 100%. The perfo...

Journal of Soft Computing Exploration, 2021

Mushrooms are plants that are widely consumed by the general public, but not all mushrooms can be consumed directly, because the types of mushrooms are feasible and it is still too difficult to distinguish, then there are several ways to identify fungi, namely by means of morphology. The morphology referred to in this paper is the morphology of fungi which includes color, habitat, class, and others. We got the morphology of this mushroom from a datasets we get from UCI Machine Learning with the 23 atribut that we use in the program. In determining the classification of this fungus we use the Naive Bayes algorithm which produces an accuracy of around 90,2% which we then improve again so that it reaches 100% accuracy using the K-Nearest Neighbors algorithm. Furthermore, in this case to prove accuracy that we had before, we use calculation accuracy with confusion matrix to show it the accuracy of classification poisonous or edible mushroom.

International Journal of Advanced Trends in Computer Science and Engineering, 2020

Mushroom possesses potential benefits in terms of antioxidants, which are essential to the body. There are various health benefits one can get on consuming mushrooms, but not all types are said to be editable. There are types of mushrooms that are poisonous. The research aims to classify which among the mushroom are edible and poisonous based on given attributes such as odor, shape, texture, and color. The data set is originally donated from the Machine Learning Repository Department of the University of California in Irvine. This research focuses on analyzing the performance of the classification algorithms in mushroom poisonous detection. The researchers utilized the Knowledge Discovery in Databases processes in performing pattern extraction and performance analysis of algorithms. The decision tree method that is based on the entropy and information gain calculations has the highest computed accuracy result compare to other classification algorithms.