Toward lightweight intrusion detection systems using the optimal and efficient feature pairs of the Bot-IoT 2018 dataset

Intrusion detection systems play a vital role in traffic flow monitoring on Internet of Things networks by providing a secure network traffic environment and blocking unwanted traffic packets. Various intrusion detection systems approaches have been proposed previously based on data mining, fuzzy techniques, genetic, neurogenetic, particle swarm intelligence, rough sets, and conventional machine learning. However, these methods are not energy efficient and do not perform accurately due to the inappropriate feature selection or the use of full features of datasets. In general, datasets contain more than 10 features. Any machine learning–based lightweight intrusion detection systems trained with full features turn into an inefficient and heavyweight intrusion detection systems. This case challenges Internet of Things networks that suffer from power efficiency problems. Therefore, lightweight (energy-efficient), accurate, and high-performance intrusion detection systems are paramount instead of inefficient and heavyweight intrusion detection systems. To address these challenges, a new approach that can help to determine the most effective and optimal feature pairs of datasets which enable the development of lightweight intrusion detection systems was proposed. For this purpose, 10 machine learning algorithms and the recent BoT-IoT (2018) dataset were selected. Twelve best features recommended by the developers of this dataset were used in this study. Sixty-six unique feature pairs were generated from the 12 best features. Next, 10 full-feature-based intrusion detection systems were developed by training the 10 machine learning algorithms with the 12 full features. Similarly, 660 feature-pair-based lightweight intrusion detection systems were developed by training the 10 machine learning algorithms via each feature pair out of the 66 feature pairs. Moreover, the 10 intrusion detection systems trained with 12 best features and the 660 intrusion detection systems trained via 66 feature pairs were compared to each other based on the machine learning algorithmic groups. Then, the feature-pair-based lightweight intrusion detection systems that achieved the accuracy level of the 10 full-feature-based intrusion detection systems were selected. This way, the optimal and efficient feature pairs and the lightweight intrusion detection systems were determined. The most lightweight intrusion detection systems achieved more than 90% detection accuracy.

Time-domain Feature and Ensemble Model based Classification of EMG Signals for Hand Gesture Recognition

The electromyography (EMG) signal is the electrical current generated in muscles due to the inter-change of ions during their contractions. It has many applications in clinical diagnostics and the biomedical field. This paper has experimented with various ensemble algorithms and time-domain features to classify eight types of hand gestures. To train and test the machine learning models, we have extracted eight types of time-domain features from the raw EMG signals, such as integrated EMG (IEMG), variance, mean absolute value (MAV), modified mean absolute value type 1, waveform length, root mean square, average amplitude change, and difference absolute standard deviation value. The ensemble machine learning models are based on stacking, bagging, and gradient boosting. We have used four different-sized training sets to evaluate the performance of these classifiers. From the performance evaluation, we have identified the XG-Boost (gblinear) classifier with the IEMG feature as the best classifier-feature pair. The proposed classifier-feature pair has given better performance with a classification accuracy of 98.33% and a processing time of 5.67 μs for one vector than the existing extended associative memory-MAV classifier-feature pair.

MACMIC Reveals Dual Role of CTCF in Epigenetic Regulation of Cell Identity Genes

ABSTRACTNumerous studies of relationship between epigenomic features have focused on their strong correlation across the genome, likely because such relationship can be easily identified by many established methods for correlation analysis. However, two features with little correlation may still colocalize at many genomic sites to implement important functions. There is no bioinformatic tool for researchers to specifically identify such feature pair. Here, we develop a method to identify feature pair in which two features have maximal colocalization but minimal correlation (MACMIC) across the genome. By MACMIC analysis of 3,385 feature pairs in 15 cell types, we reveal a dual role of CTCF in epigenetic regulation of cell identity genes. Although super-enhancers are associated with activation of target genes, only a subset of super-enhancers colocalized with CTCF regulate cell identity genes. At super-enhancers colocalized with CTCF, the CTCF is required for the active marker H3K27ac in cell type requiring the activation, and also required for the repressive marker H3K27me3 in other cell types requiring the repression. Our work demonstrates the biological utility of the MACMIC analysis and reveals a key role for CTCF in epigenetic regulation of cell identity.

Feature Pair Index Graph for Clustering

Text documents are significant arrangements of various words, while images are significant arrangements of various pixels/features. In addition, text and image data share a similar semantic structural pattern. With reference to this research, the feature pair is defined as a pair of adjacent image features. The innovative feature pair index graph (FPIG) is constructed from the unique feature pair selected, which is constructed using an inverted index structure. The constructed FPIG is helpful in clustering, classifying and retrieving the image data. The proposed FPIG method is validated against the traditional KMeans++, KMeans and Farthest First cluster methods which have the serious drawback of initial centroid selection and local optima. The FPIG method is analyzed using Iris flower image data, and the analysis yields 88% better results than Farthest First and 28.97% better results than conventional KMeans in terms of sum of squared errors. The paper also discusses the scope for further research in the proposed methodology.