Comparison Support Vector Machine and Random Forest Algorithms in Detect Diabetes

One of the diseases that are very concerning and numerous cases that occur all over the world because the impact is very significant is diabetes. Diabetes sufferers experience disorders in metabolism that identify hyperglycemia caused by no the inability of the pancreas to secrete insulin, which has an impact on death Because No functioning of other body organs. Data states in 2019 that, 433 million people were diagnosed with diabetes, and the Number is predicted to increase until peaking in 2045 will be 700 million people. This matter needs to be anticipated as soon as possible, perhaps by society, with several characteristics that occur in patients. Data on diabetes sufferers can processed with data mining that utilizes machine learning to detect diabetes. Study This will compare the Support Vector Machine and Random Forest algorithms to find accurate results. Researchers use the KDD (Knowledge Discovery in Database) model in several stages, such as data selection, preprocessing, Transformation, and Evaluation. The dataset used was sourced from the kaggel.com website; there were 768, consisting of 500 negative and 268 positive for diabetes. The SVM algorithm with a linear kernel produces a mark accuracy of 77%, Precision of 75%, Recall of 51%, and F1 score of 61%. For the Random Forest algorithm with n_estimators =100, random_state =42, results mark Accuracy 75%, precision 69%, recall 55% and F1 score 61%. The process and results state that more SVM algorithms are suitable for detecting diabetes. Models made using the Python programming language will be implemented with stremlit so you can use Web-based.