Gamified Online WEKA

Abstract

Data mining is the process of analyzing data from different perspectives and summarizing it into useful information. Data generated by different industries needs to be analyzed and summarized to help in the growth of business. Data mining applications are widely used in direct marketing, health industry, ecommerce, customer relationship management (CRM), telecommunication industry and financial sector. WEKA is one of the most commonly used open source data mining tool. Its Java API is freely available. So, it can be embedded in any java based software. It is constantly improving and new versions are being released since 2004. Its architecture is very simple and it can read data from 10 different file formats, URL of file and relational database. WEKA has more than 220 algorithms for different types of data processing (preprocess, classification, Clustering etc). Desktop based WEKA needs installation and configuration. It uses system resources and has a maximum heap size limit. All the visualizations (trees, bar charts and scatter plots) are static in WEKA and it doesn’t convey the required information. Results of the filters are displayed in a very user unfriendly way. As a solution to this problem we have created “Gamified Online WEKA”. It has improved the visualizations of WEKA. As to use it user would need to upload and store private data on server, it maintains separate and secure user accounts and stores their data for later use. With its collaborative environment, data security, data storage, interactive visualizations and all the algorithms of WEKA, “Gamified Online WEKA” is a complete data mining solution.