nFuture - IT Fist Research Lab

About the lab

Where others see a full stop, We see the possibilities.

Research Area:

1. Information Security
2. Machine Learning and Deep Learning
3. Bioinformatics and Computational Biology
4. Software Engineering
5. Blockchain Technology

Featured projects (1)

Stacked generalization (or stacking) (Wolpert, 1992) is a different way of combining multiple models, that introduces the concept of a meta learner. Although an attractive idea, it is less widely used than bagging and boosting. Unlike bagging and boosting, stacking may be (and normally is) used to combine models of different types. The procedure is as follows: 1. Split the training set into two disjoint sets. 2. Train several base learners on the first part. 3. Test the base learners on the second part. 4. Using the predictions from 3) as the inputs, and the correct responses as the outputs, train a higher level learner. Note that steps 1) to 3) are the same as cross-validation, but instead of using a winner-takes-all approach, we combine the base learners, possibly non-linearly. The Goal of this project is to implement and evaluate the stacked generalization in different areas. The effectiveness of stacking concept will be identified on which are it will better fit.

Featured research (16)

Fake news is any content or information that is false and often generated to mislead its readers in believing something which is not true. Fake news has become one of major threats that can harm someone’s reputation. It often circulates wrong or made up information about various products, events, people or entity. The deliberate making of such news is escalating drastically these days. Fake news deceives us in taking wrong decisions. Therefore, Fake News Detection has attained immense deal of interest from researchers all over the world. In this chapter, a machine learning approach has been proposed named FakeTouch starting with Natural Language Processing based concept by applying text processing, cleaning and extraction techniques. This approach aim to arrange the information to be “obeyed” into each classification model for training and tuning parameters for every model to bring out the optimized and best prediction to find out the Fake news. To evaluate the proposed framework, three use cases with three different datasets has been developed during this study. The proposed framework will also help to understand what amount of data is responsible for detecting fake news, trying to stage the linguistic differences between fake and true articles providing a visualization of the results using different visualization tools. This chapter also presents a comprehensive performance evaluation to compare different well known machine learning classifiers like Support Vector Machine, Naïve Bayes Method, Decision Tree Classifier, Random Forest, Logistic Regression as well as to develop an ensemble method (Bagging & Boosting) like XGBClassifier, Bagging Classifier of different combinations of classification models to identify which will give the best optimal results for three part of datasets. As a result, it has been found that with an appropriate set of features extracted from the texts and the headlines, XGB classifier can effectively classify fake news with very high detection rate. This framework also provides a strong baseline of an intelligent anti-fake news detector.
Social media is using rapidly for expressing user’s opinions and feelings. At present, ‘Depression’ is propagating in a cursory way which is the reason for the increasing rate of suicides even more. Recently, the status on social media can give the hints of user’s mental state along with the situation and activities that are happening with them. Different research works had done for perceiving the mental health of any user through social media which has impressive impact. That is done by analysing expressed opinions, images, sentiments, linguistic style and other activities. Variant studies had proposed variant methods or intelligent systems for detecting depression or state of mental health through social media posts. In this chapter, the posted tweets from users on Twitter will be considered in the sake of detecting depressions. Six machine learning approaches named Multinomial Naive Bayes, Support Vector Classifier, Decision Tree, Random Forest, K-Nearest Neighbor, Logistic Regression had used to differentiate the depressed and non-depressed users. Finally, Support Vector Classifier outperforms among all of the investigated and evaluated techniques with 79.90% accuracy, 75.73% precision, 77.53% recall and 76.61% f1-factor. This study can be a basement for the developers of intelligent systems in the area of users mental conditions detection.
Phishing is an alarming issue among the cybercriminals. In the last decade, online services have revolutionized the world. Due to the revolutionary transformations of web service, the reliance on the web has increased day by day. Security threats have emerged due to the increasing reliance on online orientation. There are many types of anti-phishing solutions available that have been proposed by many researchers. However, this chapter is to propose an intelligent framework to detect phishing URLs based on the optimized learning architecture scheme. Multi-layer based structures have been implemented to detect phishing URLs using Deep Neural Network (DNN), Neural Network (NN) and Stacking. These architectures are evaluated with various tuning hyper-parameters to obtain the optimized output named AntiPhishTuner. As a result, five-layer based DNN can provide accuracy of 0.95 with the minimum mean squared error (MSE) 0.30, and also a mean absolute error (MAE) 0.074 where the number of epochs was 50 and Adam optimizer as an optimizer. Using two-layer NN with AdaGard optimizer can provide accuracy of 0.95, with MSE 0.30 and MAE 0.074. NN provides these results with 150 epochs. Stack generalization can reach maximum accuracy 0.97 in binary classification with MAE 2.1. This chapter can provide a better lead to researchers and anti-phishing tools developers to make an initial decision about the approach that should be followed for further extension.
Alongside the recognition of the android operating system (OS), android malware is on the increase. Cybercriminals are using different techniques to develop malware for android devices. In addition, malware authors are trying to make malicious android applications that severely undermine the potential of traditional malware detectors. The key purpose of the chapter is to analyze and have a different appearance at various techniques of Android malware detection in a variety of research articles. However, this chapter presents an analysis of varied android malware detection approaches and comparing them to supported various parameters like detection technique, analysis method, features extracted and so on. The experiments are based on substantial malware datasets, evaluation parameters and this study employ a wide variety of machine learning techniques, including decision trees and random forests, support vector machines, logistic model trees, and artificial neural networks, also Deep learning techniques. It is a comparative analysis that should be useful in this field for researchers. The analysis shows, based on simple criteria, the similarities and differences in essential published research in addition to the accuracy. Thus, this chapter aims to study various android malware detection techniques and to identify plausible research directions. The findings showed that machine learning, with greater detection accuracy, is a more promising method. In order to achieve improved accuracy, future researchers can pursue a deep learning approach with the use of a large dataset.
Android Malware has grown dramatically day by day because of the rising trends of android operating based smartphones. It has become the main attraction point by attackers now-a-days. Thus, android malware detection has become a major field of investigation among the researchers and academicians who are working with in the field of cyber security. As there are lots of research works have done already, it is still major matter of concern to improve the anti-malware tools. In addition, during the development of anti-malware framework the features of android malware plays the major role. During this study, an important features identification and selection technique has been proposed named IFIFDroid and evaluated which is based on wrapper method. However, the proposed approach can minimize the number of features which helps to machine learning (ML) techniques to learn from less features but perform better. It’s found that IFIFDroid can ranking features based on the capacity of individual ML algorithms and comparatively provide better result than existing wrapper method. IFIFDroid proves that there is still way to improve the features selection scheme and provide a strong basement of minimizing the power, execution time during the training by ML algorithms. Though if there is less features to fit without losing accuracy then it will minimize the processing resources as well.

Lab head

Sheikh Shah Mohammad Motiur Rahman
  • Département Informatique des Systèmes Complexes (DISC)
About Sheikh Shah Mohammad Motiur Rahman
  • Thirsty to learn. Curious to discover thyself. Passionate to draw new conclusions.

Members (8)

Takia Islam
  • Daffodil International University
Md Fahimuzzman Sohan
  • Daffodil International University
Abu Bakkar Siddikk
  • Daffodil International University
Md. Fahim Muntasir
  • Daffodil International University
Md. Omar Faruque Khan Russel
  • University of Hertfordshire
Rifat Jahan Lia
  • Daffodil International University
Md. Mahfuzur Rahman
  • Daffodil International University
Saiful Alam
  • Daffodil International University