WEKA for Reducing High -Dimensional Big Text Data

Kotonko Lumanga Manga Tresor, Professor Xu Dhe zi

Dimension Reduction; J48; WEKA; MATLAB.

In the current era, data usually has a high volume, variety, velocity, and veracity, these are known as 4 V’s of Big Data. Social media is considered as one of the main causes of Big Data which get the 4 V’s of Big Data beside that it has high dimensionality. To manipulate Big Data efficiently; its dimensionality should be decreased. Reducing dimensionality converts the data with high dimensionality into an expressive representation of data with lower dimensions. This research work deals with efficient Dimension Reduction processes to reduce the original dimension aimed at improving the speed of data mining. Spam-WEKA dataset; which entails twitter user information. The modified J48 classifier is applied to reduce the dimension of the data thereby increasing the accuracy of data mining. The data mining tool WEKA is used as an API of MATLAB to generate the J48 classifiers. Experimental results indicated a significant improvement over the existing J48algorithm

10.22161/ijaers.5.11.10