In: Computer Science
You have decided to use J48 as a classifier in Weka for your data set. After your analysis, you have found that the accuracy of J48 for your data set is greater than that of ZeroR, but less than the accuracy of OneR. Should you continue to use J48 as a classifier for your data set? Why or why not?
Weka is a software which is used to process bigdata , analyzed it using different machine learning algorithms. There are several types of classifiers available in Weka which is an open source software. Classification is an important task in machine learning applications. It is used to classify the given data based on several features and characteristics which is subjected to pre-defined set of data.
classifiers mentioned in the question are ZeroR, J48, OneR
ZeroR is the simplest method of classification which has no predictability power. It just predict the majority from the available dataset. It just depends on the frequency of target.
OneR stands for one rule , simplest and most effective method of classification. It generates one rule for each predictor in the data. It uses minimum error attribute for prediction. It performs well in most of the dataset.It is simple for humans to interpret.
J48 is a simple type of classifer to make a decission tree. It is a type of statistical classifiers. It is also a predictive machine learning model which calculates result of a new sample based on various attribute values from the available data.
The main purpose of classification is to produce accurate results. However , accuracy of classification is strongly affected by the nature of dataset. The Accuracy rate of a classifying model is the percentage of test set samples that are correctly classified by the model. Out of these OneR produces more accurate results ,So we can select OneR as the classification model for better classification. It doesn't mention the nature of dataset, because different models are suited and designed for different types of dataset. So just continue with the best model.