• Muralidhara S

Abstract: Anomaly Detection on Facebook images plays a key role on identification of anomalous or behavioral analysis of Facebook. It targets detection of quality, violence, perfection of the images of the Facebook. Using specific Artificial Neural network algorithm of the machine learning detection of the anomalous images is performed by the project. Convolution neural network algorithm with 5 layers has been selected in this project to detect some of the anomalous images from a large dataset of the images downloaded from the Facebook. The total dataset of the images is divided into 3 parts such as train dataset, validation dataset, and Test dataset. The accuracy of the algorithm depends on the volume of the training dataset. The larger training dataset gives more accurate results compare to the smaller dataset. Kernel matrix has been formed from the RGB pixel of the images downloaded from Facebook. Each image have kernel matrix. From the kernel matrix algorithm is being trained. The objective of the project is to detect quality; violence and perfection of downloaded images from the Facebook which intern detect the activities and behavior of the user. Facebook has its own immune system to safeguard its users from unwanted malicious content. Despite the immune system deployed by Facebook, unwanted spam, phishing, and other malicious content continues to exist on Facebook. Objective is to detection of malicious content on Facebook .The goal is to decide for each user whether the account is compromised or not.

Artificial neural Network, Machine Learning, Social Networks

"ANAMOLY DETECTION USING CNN ALGORITHM", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 6, page no.629-631, June 2019, Available :http://www.jetir.org/papers/JETIR1906P91.pdf

"ANAMOLY DETECTION USING CNN ALGORITHM", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 6, page no. pp629-631, June 2019, Available at : http://www.jetir.org/papers/JETIR1906P91.pdf