Web based clustering tool using K-MEAN++ algorithm / Muhammad Nur Syazwanie Aznan

Cluster analysis is one of the data mining task that are widely used in many area to extracting, grouping data with similar attribute in order to uncover the hidden pattern and meaning in the data. Therefore, many mathematicians and programmers have discovered many techniques and developed tools to help society to do their researches. But until today, there are only a few tools that provide the web based platform as their tools for example MATLAB MWS and Clustvis. Even though, there is a tool for clustering, some of this tool required an expert knowledge in clustering in order to understand the results. Which is why this project objective is to develop a web based clustering tool using K-MEAN++ algorithm. This project will use the rapid application development (RAD) methodology since this are the most suitable method for developing the system. The first phase in rapid application development is requirement and planning, this is where the problem statement, objective, scope, significance of this project are defined. The next phase is design, flowchart and use case diagram of this project will be design and follow accordingly. The other phases are construction where this project engine are developed using the JAVASCRIPT language and HTML for user interfaces. For testing, this project used a dummy data created with forty instances and ten attributes and Iris data contain of one hundred and fifty instances and 4 attributes. Although the tool has some limitation, but this tool are successfully tested with this two data and manage to calculate and shows the cluster with silhouette score. To sum up, these projects successfully manage to achieve the objective which is developing the web based clustering tool and all functionality are working properly.