Mobile forensics analysis is the sub-domain of digital forensics, which addresses solving the minor technology misuse cases to substantial international digital crime cases. Mobile forensic refers to the acquisition of data and analysis of the artifacts collected from the mobile devices. Mobile phones are used as a means of communication and have evolved to a mini-portable computer having the advanced communication capabilities. New threats and challenges are being faced in the domain of mobile forensics by every passing year. In this paper, we review forensic analysis techniques for android phones and perform a critical analysis of the recent trends and techniques in the field of mobile forensics. We provide a comprehensive overview to the current state-of-the-art in this area. We identify new methodologies, tools and techniques which are successfully being used for forensic investigations of the mobile phones. With the help of this analysis we identify the key challenges and knowledge gaps for potential future research work.

With the development of wireless communications, several studies have been performed on Location based Services due to their numerous applications. Amongst those recommendations, Travel Planning and Recommendations are few of the active topics. When it comes to movement patterns and mobility, human beings are restricted in motion due to social, financial and geographical constraints. Using check-in data from a former location based social network namely Gowalla and airport flights and route data from openflights.org, authors aim to extract a graph model from time series data of user check-ins. In this study, authors have identified patterns of location (latitude, longitude) visits using directed weighted graph. In addition to it, we have mapped airports to identified maps using latitude, longitude and used routes. This has helped to identify nearest airport routes probably used by Gowalla users. Hence, by mapping two heterogeneous graphs i.e. the Gowalla check-in data and openflights.org airport and flights data, we have tried to extract the most commonly used airport routes travelled by Gowalla users.

The major focus in this paper is to get the factors that shows the significance in predicting the risks of virus of hepatitis-C. 2 datasets were used for this purpose the first one is gathered from UCI Repository and the second one is taken from Zahid Medical Centre with the help of Dr. Abdul Fateh. There are nineteen features and a class feature with classification in binary. The first data set that is gathered from UCI repository contains 155 records with missing values in most of them in order to reduce this technique of normalization is applied. Now for qualitative approaches for data reduction as well as quantitative the binary logistic regression is used. The first result gathered from the Zahid Medical Centre gave us 58% accuracy result using these techniques. And second result using these procedures produced about 90% accurate classification. Our approach gives good classification rate only by using total 37% fields.

Named Entity Recognition is an essential module of Information Extraction in the field of bio-medical and diseases are one of the most important sector to study in the medical field, but since the amount of incessantly updated information on diseases is huge and is merely accessible in the form of published journals or articles. An efficient Named Entity Recognizer is needed to extract diseases directly from the input given in the form of articles and to annotate the extracted terms with the knowledge base. The Named Entity Recognizer techniques must first identify the targeted terms. Though biomedical articles often consist of proper nouns recently prepared by the authors, and dictionaries which are conventional methods based on domain specific cannot identify such unidentified words. This study will identify a better and efficient Information extraction system which will extract diseases from the given biomedical text using techniques such as dictionary based and machine learning based (K-nearest neighbor and Naïve Bayes) techniques. The efficiency of both techniques and both algorithms have been measured through confusion matrix and machine learning approach more specifically K-nearest neighbor has been found more proficient as compared to other techniques.