<![CDATA[In this article, the Naïve Bayes Classifier is employed to detect fraud in automobile insurance. The Naïve Bayes classier is a simple probabilistic method based on the Bayes theorem. The data used in this article is determined from databricks.com which consists of 40 attributes and 1000 entries. The target attribute that will be predicted consists of two categories,” yes" or "no", which inform whether there is a fraud or not. The Data is split into training and testing with suitable proportions. Based on training data, the Naïve Bayes Classifier is applied to the testing data and returns the predictions data. Then, the prediction data is compared with the actual data to see the performance of the method. The result shows that the Naïve Bayes Classifier gives a good result to predict the insurance fraud with 78% accuracy, 67% precision, 3% of recall, and 6% of F1 score for “Yes”]]>
Copyrights © 2022
