Md. Kamrul Hasan

A cohesion-based friend-recommendation system

Social network sites have attracted millions of users with the social revolution in Web 2.0. A social network is composed of communities of individuals or organizations that are connected by common interests. In the social network sites, a user can register other users as friends and enjoy communication. In such a scenario, often users face problems in finding their appropriate friends using typical friend-of-friend-recommendation methods. In this paper, we propose a friend-recommendation system based on cohesion. We analyze the cohesive subgroup on an augmented network formed by the physical connection network with the information of common interests and interactions. All the users in a cohesive subgroup should be friends to each other. We validated our idea on a small friend network.

Activity recognition of a badminton game through accelerometer and gyroscope

The scope for doing physical exercises in daily life is declining day by day. But, the importance of human physical exercise for a healthy life, remains the same. It is necessary to generate a solution to simulate the outdoor experience of physical exercises and sports inside our home. In this paper, we propose an idea of recognizing the activities of a badminton game which has the potential to be useful in simulating the Badminton Sport. We have used motion sensors (e.g. Accelerometer, Gyroscope) to recognize different activities like, serve, smash, backhand, forehand, return etc. We have collected data from a large set of users and labeled their data over several instances. We have applied the K-Nearest Neighbors (k-NN) and Support Vector Machines (SVM) classifiers to recognize those activities. Existing approaches (e.g. Microsoft Xbox 360) used vision based techniques to recognize activities and use it in simulated games but we are using sensor based approach. Vision based approaches have some limitations such as slow rate of data, illumination constraints, occluded backgrounds etc. Our approach gives a low cost solution with a classification technique which is faster. The experimental result shows a decent recognition rate.

EXPRESit: Interactive messaging via gestures

Text communication or messaging is devoid of emotional expression, gesture and intonation. As a result text communication has become a robotic conversation between human beings. People uses various applications like, Facebook Messenger, Viber, WhatsApp etc. for the purpose of messaging, where they follow the traditional system of writing messages and sending to the receivers. This paper, presents a new idea of messaging through gestures which is one of the primary elements of interaction. The message will be produced through the gestures that the sender will make. If, the sender shows a ‘thumbs up’ gesture, a text or emoticon is generated and sent to the receiver. One of the best features of our system is, it can generate Bengali messages. The system can recognize ‘thumbs up’ gesture which is converted to “???”, means ‘Well Done’ and sent to the receiver. We have evaluated our designed prototype and got satisfactory results.

Linear regression-based feature selection for microarray data classification

Predicting the class of gene expression profiles helps improve the diagnosis and treatment of diseases. Analysing huge gene expression data otherwise known as microarray data is complicated due to its high dimensionality. Hence the traditional classifiers do not perform well where the number of features far exceeds the number of samples. A good set of features help classifiers to classify the dataset efficiently. Moreover, a manageable set of features is also desirable for the biologist for further analysis. In this paper, we have proposed a linear regression-based feature selection method for selecting discriminative features. Our main focus is to classify the dataset more accurately using less number of features than other traditional feature selection methods. Our method has been compared with several other methods and in almost every case the classification accuracy is higher using less number of features than the other popular feature selection methods.

Better user recommendations using enhancing software development process repository

Reusing previously completed software repository to enhance the development process is a common phenomenon. If developers get suggestions from the existing projects they might be benefited a lot what they eventually expect while coding. The strategies available in this field have been rapidly changing day by day. There are a number of efforts that have been focusing on mining process and constructing repository. Some of them have emphasized on the web based code searching while others have integrated web based code searching in their customized tool. But web based approaches have inefficiency especially in building repository on which they apply mining technologies. To search the code snippets in response to the user query we need an enriched repository with a better representation and abstraction. To ensure that repository before mining process we have developed a concept based on Enhancing Software Development Process (ESDP). In ESDP approach multiple sources of codes from both online and offline storages are considered to construct the central repository with XML representation and applied mining techniques in the client side. The respective evaluation shows that ESDP approach works much better in response time and performance than many other existing approaches available today.