Venkatanareshbabu Kuppili

SM-RuleMiner

Diabetes is a major health challenge around the world. Existing rule-based classification systems have been widely used for diabetes diagnosis, even though they must overcome the challenge of producing a comprehensive optimal ruleset while balancing accuracy, sensitivity and specificity values. To resolve this drawback, in this paper, a Spider Monkey Optimization-based rule miner (SM-RuleMiner) has been proposed for diabetes classification. A novel fitness function has also been incorporated into SM-RuleMiner to generate a comprehensive optimal ruleset while balancing accuracy, sensitivity and specificity. The proposed rule-miner is compared against three rule-based algorithms, namely ID3, C4.5 and CART, along with several meta-heuristic-based rule mining algorithms, on the Pima Indians Diabetes dataset using 10-fold cross validation. It has been observed that the proposed rule miner outperforms several well-known algorithms in terms of average classification accuracy and average sensitivity. Moreover, the proposed rule miner outperformed the other algorithms in terms of mean rule length and mean ruleset size.

Diabetes Classification using Radial Basis Function Network by Combining Cluster Validity Index and BAT Optimization with Novel Fitness Function

Diabetes is one of the foremost causes for the increase in mortality among children and adults in recent years. Classification systems are being used by doctors to analyse and diagnose the medical data. Radial basis function neural networks are more attractive for classification of diseases, especially in diabetes classification, because of it’s non iterative nature. Radial basis function neural networks are four layer feed forward neural network with input layer, pattern layer, summation layer and the decision layer respectively. The size of the pattern layer increases on par with training data set size. Though various attempts have made to solve this issue by clustering input data using different clustering algorithms like k-means, k-medoids, and SOFM etc. However main difficulty of determining the optimal number of neurons in the pattern layer remain unsolved. In this paper, we present a new model based on cluster validity index with radial basis neural network for classification of diabetic patients data. We employ cluster validity index in class by class fashion for determining the optimal number of neurons in pattern layer. A new convex fitness function has also been designed for bat inspired optimization algorithm to identify the weights between summation layer and pattern layer. The proposed model for radial basis function neural network is tested on Pima Indians Diabetes data set and synthetic data sets. Experimental results proved that our approach performs better in terms of accuracy, sensitivity, specificity, classification time, training time, network complexity and computational time compared to conventional radial basis function neural network. It is also proved that proposed model performs better compared to familiar classifiers namely probabilistic neural network, feed forward neural network, cascade forward network, time delay network, artificial immuine system and GINI classifier.

RST-BatMiner: A fuzzy rule miner integrating rough set feature selection and Bat optimization for detection of diabetes disease

Abstract Fuzzy classification rules are more interpretable and cope better with pervasive uncertainty and vagueness with respect to crisp rules. Because of this fact, fuzzy classification rules are extensively used in classification and decision support systems for disease diagnosis. But, most of the rule mining techniques failed to generate accurate and comprehensive fuzzy rules. This paper presents a hybrid decision support system based on Rough Set Theory (RST) and Bat optimization Algorithm (BA) called RST-BatMiner. It consists of two stages. In the first stage, redundant features have been removed from the data set through RST-based QUICK-REDUCT approach. In the second stage, for each class BA is invoked to generate fuzzy rules by minimizing proposed fitness function. Further, an Ada-Boosting technique is applied to the rules generated by BA to increase the accuracy rate of generated fuzzy rules. Moreover, to generate comprehensive fuzzy rules, a new ≠ (not equal) operator along with = (equal) operator is introduced into BA encoding scheme. The proposed RST-BatMiner builds consolidated fuzzy ruleset by learning the rules associated with each class separately. The proposed RST-BatMiner is experimented on six bench-mark datasets namely Pima Indians Diabetes, Wisconsin Breast Cancer, Cleveland Heart disease, iris, wine and glass, in order to validate its generalization capability. These experimental results show that except for wine dataset the proposed RST-BatMiner yields high accuracy and comprehensible ruleset when compared to other state-of-the-art bio-inspired based fuzzy rule miners and Fuzzy Rule Based Classification Systems (FRBCS) in the literature. In the case of wine dataset, the proposed RST-BatMiner yields second highest accuracy along with comprehensible ruleset.

Heterogeneous node split measure for decision tree construction

A new heterogeneous node split measure (HSM) has been proposed in this paper for decision tree construction. The split measure HSM is derived from quasilinear mean of information gain. This helps in including proportionalities of class values from the sub-partitions and the entire dataset at the same time. This results in acquiring more information at the split point, which produces compact decision trees. Comparative performance evaluation of HSM on benchmark datasets with the well known node splitting measures Gini-index and Gain ratio shows that HSM is capable of generating decision trees which are lesser in height. The classification accuracy is also far superior and the computational time is also less using HSM as the split measure.

Deep learning strategy for accurate carotid intima-media thickness measurement: An ultrasound study on Japanese diabetic cohort

MOTIVATION The carotid intima-media thickness (cIMT) is an important biomarker for cardiovascular diseases and stroke monitoring. This study presents an intelligence-based, novel, robust, and clinically-strong strategy that uses a combination of deep-learning (DL) and machine-learning (ML) paradigms. METHODOLOGY A two-stage DL-based system (a class of AtheroEdge™ systems) was proposed for cIMT measurements. Stage I consisted of a convolution layer-based encoder for feature extraction and a fully convolutional network-based decoder for image segmentation. This stage generated the raw inner lumen borders and raw outer interadventitial borders. To smooth these borders, the DL system used a cascaded stage II that consisted of ML-based regression. The final outputs were the far wall lumen-intima (LI) and media-adventitia (MA) borders which were used for cIMT measurements. There were two sets of gold standards during the DL design, therefore two sets of DL systems (DL1 and DL2) were derived. RESULTS A total of 396 B-mode ultrasound images of the right and left common carotid artery were used from 203 patients (Institutional Review Board approved, Toho University, Japan). For the test set, the cIMT error for the DL1 and DL2 systems with respect to the gold standard was 0.126 ± 0.134 and 0.124 ± 0.100 mm, respectively. The corresponding LI error for the DL1 and DL2 systems was 0.077 ± 0.057 and 0.077 ± 0.049 mm, respectively, while the corresponding MA error for DL1 and DL2 was 0.113 ± 0.105 and 0.109 ± 0.088 mm, respectively. The results showed up to 20% improvement in cIMT readings for the DL system compared to the sonographers readings. Four statistical tests were conducted to evaluate reliability, stability, and statistical significance. CONCLUSION The results showed that the performance of the DL-based approach was superior to the nonintelligence-based conventional methods that use spatial intensities alone. The DL system can be used for stroke risk assessment during routine or clinical trial modes.

Symtosis: A liver ultrasound tissue characterization and risk stratification in optimized deep learning paradigm

Background and Objective Fatty Liver Disease (FLD) - a disease caused by deposition of fat in liver cells, is predecessor to terminal diseases such as liver cancer. The machine learning (ML) techniques applied for FLD detection and risk stratification using ultrasound (US) have limitations in computing tissue characterization features, thereby limiting the accuracy. Methods Under the class of Symtosis for FLD detection and risk stratification, this study presents a Deep Learning (DL)-based paradigm that computes nearly seven million weights per image when passed through a 22 layered neural network during the cross-validation (training and testing) paradigm. The DL architecture consists of cascaded layers of operations such as: convolution, pooling, rectified linear unit, dropout and a special block called inception model that provides speed and efficiency. All data analysis is performed in optimized tissue region, obtained by removing background information. We benchmark the DL system against the conventional ML protocols: support vector machine (SVM) and extreme learning machine (ELM). Results The liver US data consists of 63 patients (27 normal/36 abnormal). Using the K10 cross-validation protocol (90% training and 10% testing), the detection and risk stratification accuracies are: 82%, 92% and 100% for SVM, ELM and DL systems, respectively. The corresponding area under the curve is: 0.79, 0.92 and 1.0, respectively. We further validate our DL system using two class biometric facial data that yields an accuracy of 99%. Conclusion DL system shows a superior performance for liver detection and risk stratification compared to conventional machine learning systems: SVM and ELM.

PSO-RBFNN: A PSO-Based Clustering Approach for RBFNN Design to Classify Disease Data

The Radial Basis Function Neural Networks (RBFNNs) are non-iterative in nature so they are attractive for disease classification. These are four layer networks with input, hidden, output and decision layers. The RBFNNs require single iteration for training the network. On the other side, it suffers from growing hidden layer size on par with training dataset. Though various attempts have been made to solve this issue by clustering the input data. But, in a given dataset estimating the optimal number of clusters is unknown and also it involves more computational time. Hence, to address this problem in this paper, a Particle Swarm Optimization (PSO)-based clustering methodology has been proposed. In this context, we introduce a measure in the objective function of PSO, which allows us to measure the quality of wide range of clusters without prior information. Next, this PSO-based clustering methodology yields a set of High-Performance Cluster Centers (HPCCs). The proposed method experimented on three medical datasets. The experimental results indicate that the proposed method outperforms the competing approaches.

An Efficient Load Balancing of Gateways Using Improved Shuffled Frog Leaping Algorithm and Novel Fitness Function for WSNs

Energy consumption is one of the important factors in wireless sensor networks (WSNs) design. As energy is a limited resource, energy consumption problem in WSNs has become a fast growing problem, and there is a need of efficient and robust algorithms for load balancing in WSNs. This energy is needed for sensor nodes operations. In order to maximize the network lifetime, energy consumption should be optimized. In cluster-based WSNs, cluster heads or gateways perform activities, such as data collection from its member nodes, data aggregation, and data exchange with the base station. Hence, load balancing of gateways in WSNs is one of the crucial and challenging tasks to maximize network lifetime. In order to address this problem, in this paper, shuffled frog leaping algorithm (SFLA) is improved by suitably modifying the frog’s population generation and off-spring generation phases in SFLA and by introducing a transfer phase. A novel fitness function is also designed to evaluate the quality of the solutions produced by the improved SFLA. We performed extensive simulations of the proposed load balancing algorithm in terms of various performance parameters. The experimental results are encouraging and demonstrated the efficiency of the proposed algorithm.

Automatic disease diagnosis using optimised weightless neural networks for low-power wearable devices

Low-power wearable devices for disease diagnosis are used at anytime and anywhere. These are non-invasive and pain-free for the better quality of life. However, these devices are resource constrained in terms of memory and processing capability. Memory constraint allows these devices to store a limited number of patterns and processing constraint provides delayed response. It is a challenging task to design a robust classification system under above constraints with high accuracy. In this Letter, to resolve this problem, a novel architecture for weightless neural networks (WNNs) has been proposed. It uses variable sized random access memories to optimise the memory usage and a modified binary TRIE data structure for reducing the test time. In addition, a bio-inspired-based genetic algorithm has been employed to improve the accuracy. The proposed architecture is experimented on various disease datasets using its software and hardware realisations. The experimental results prove that the proposed architecture achieves better performance in terms of accuracy, memory saving and test time as compared to standard WNNs. It also outperforms in terms of accuracy as compared to conventional neural network-based classifiers. The proposed architecture is a powerful part of most of the low-power wearable devices for the solution of memory, accuracy and time issues.

An efficient Concealed Information Test: EEG feature extraction and ensemble classification for lie identification

EEG-based lie detectors have become popular over polygraphs because it cannot be controlled by human intentions. Various studies have performed “Guilty Knowledge Test” or “Concealed Information Test” by creating a mock crime scenario to identify changes in brain potential. In this study, an individual’s behavior during lying is analyzed and a new scenario is developed for “Concealed Information Test.” This work involves a mock crime scenario using an EEG acquisition device for 10 participants. Data acquisition has been performed by placing 16 electrodes on the subjects’ scalp. For this experiment, the subject has to recognize faces of some known and unknown personalities among 10 images flashed. These images behave as stimulus for the subject which generate corresponding brain responses. Various feature extraction approaches such as statistical, time domain, frequency domain and time–frequency domain are applied to the 16- channel EEG data. For classifying a subject as guilty or innocent, five classifiers have been applied on subject-wise EEG data. Moreover, the classifiers’ ranking is considered based on the performance of classifiers. An ensemble framework is developed by aggregating the results of the best three classifiers out of the tested five classifiers. The classifiers’ results are aggregated using a weighted voting approach and have been compared with popular conventional approaches using various classification performance measures. Results present a comparative performance of different feature extraction approaches and classifiers using subject-wise single-trial EEG data. The wavelet approach performs better for EEG data of most of the subjects. A comparison between base classifiers and ensemble framework is provided with the ensemble approach outperforming over the base classifiers. Further proposed framework is compared with some existing approaches, and a highest accuracy of 92.4% has been achieved.