@thesis{thesis,
		author={Deris  and UBAIDILLAH },
		title ={KLASIFIKASI POLA DATA FORENSIK WHATSAPP PADA SMARTPHONE ANDROID MENGGUNAKAN METODE SUPPORT VECTOR MACHINE (SVM)},
		year={2018},
		url={https://repository.unsri.ac.id/10065/},
		abstract={WhatsApp have reached more than a billion users, making WhatsApp one of the most widely used private mobile messaging applications. Beside as tool of communication, WhatsApp messages can also be used as digital evidence to assist the criminal investigation process. Artifacts left on WhatsApp messages are analyzed forensically or correlated one by one. In the conventional method, the analysis and classification of forensic data is done manually. Therefore, in this study will use machine learning so that the investigation process can be carried out effectively and efficiently. One machine learning method that is reliable in classification and minimizing errors is Support Vector Machine (SVM). This study uses four types of kernerls namely: RBF, Linear, Sigmoid and Polynomial. Evaluation of the results of the classification of the best parameter values was obtained at a data ratio of 80%: 20%, iterations = 200, C = 1 and ? = 0.1. The results showed that the Linear SVM kernel had the highest accuracy value of 98.3% and the lowest accuracy value in the Polynomial kernel was 89.6%. In addition, the AUC value was 99.5%, which was statistically very good.}
		}