| dc.description.abstract | Nowadays, the term Data Mining (DM) and Machine Learning (ML) have been widely known over the world and also widely applied in many fields, for instance: Healthcare, Banking, Telecommunications, Education, Entertainment, etc. Therefore, a huge numbers of applications which support users to learn and research about Data Mining and Machine Learning has been dramatically increased in recent years. One of those are Waikato Environment for Knowledge Analysis (WEKA), an open source software developed by using Java provides users with tools for data preprocessing, data visualization and implementation of several Machine Learning algorithms, so that users can develop machine learning techniques and apply it into real problems.
However, Weka still have disadvantage side, such as it is an offline application, and in classification part, it does not have a chart to compare two or more algorithms. Hence, in this thesis, we research about algorithms and then compare them. Moreover, with demand of using it conveniently, we develop the Weka Application to become the Web Weka Application. This research will give us in-depth knowledge about how to preprocess data, classify data with three algorithms: Naïve Bayes (NB), Decision Tree (specifically J48) and sequential minimal optimization (SMO). The advantages and disadvantages of each algorithm is also discussed in this thesis. Furthermore, this thesis includes instructions about using Weka API in Java to preprocess data as well as classify data and how to draw chart by using JavaScript (JS) chart library to point out the difference between three algorithms. | en_US |
