Siby Abraham

Particle swarm optimisation based Diophantine equation solver

The paper introduces particle swarm optimisation as a viable strategy to find numerical solution of Diophantine equation, for which there exists no general method of finding solutions. The proposed methodology uses a population of integer particles. The candidate solutions in the feasible space are optimised to have better positions through particle best and global best positions. The methodology, which follows fully connected neighbourhood topology, can offer many solutions of such equations.

Evolution Induced Secondary Immunity: An Artificial Immune System Based Intrusion Detection System

The analogy between immune systems and intrusion detection systems encourage the use of artificial immune systems for anomaly detection in computer networks. This paper describes a technique of applying artificial immune system along with genetic algorithm to develop an intrusion detection system. Far from developing primary immune response, as most of the related works do, it attempts to evolve this primary immune response to a secondary immune response using the concept of memory cells prevalent in natural immune systems. A genetic algorithm using genetic operators- selection, cloning, crossover and mutation- facilitates this. Memory cells formed enable faster detection of already encountered attacks. These memory cells, being highly random in nature, are dependent on the evolution of the detectors and guarantee greater immunity from anomalies and attacks. The fact that the whole procedure is enveloped in the concepts of approximate binding and memory cells of lightweight of natural immune systems makes this system reliable, robust and quick responding.

A Facial Caricature Generation system using Adaptive Thresholding

Automatic Facial Caricature Generation involves extracting the feature points and emphasizing the distinctive features of a particular face. The input face is compared to an “Average Face” based on their respective facial distance parameters. The deviations are then scaled by an “Exaggeration Rate”, thereby elevating the peculiarity of the input face. A novel approach of Adaptive Thresholding has been used for the extraction of feature points from the input face to manage the non-uniform illuminations of the input face.

Protein Sequence Classification Based on N-Gram and K-Nearest Neighbor Algorithm

The paper proposes classification of protein sequences using K-Nearest Neighbor (KNN) algorithm. Motif extraction method N-gram is used to encode biological sequences into feature vectors. The N-gram generated is represented using Boolean data representation technique. The experiments are conducted on dataset consisting of 717 sequences unequally distributed into seven classes with a sequence identity of 25 %. The number of neighbors in the KNN classifier is varied from 3, 5, 7, 9, 11, 13 and 15. Euclidean distance and Cosine coefficient similarity measures are used for determining nearest neighbors. The experimental results revealed that the procedure with Cosine measure and the number of neighbors as 15 gave the highest accuracy of 84 %. The effectiveness of the proposed method is also shown by comparing the experimental results with those of other related methods on the same dataset.

Predicting 3D structure of proteins from genomic sequences: A genetic algorithm approach

The paper proposes a methodology for predicting 3D structure of proteins using genetic algorithm. It uses genomic sequences for the experimental purpose. In order to give a complete representation of known and unknown genomic sequences of similar kind, the known collection of sequences are made to evolve. The evolved sequences are subjected to offer consensus of the sequences. This consensus is used for generating a primary protein sequence, which is used for predicting the 3D structure of protein. The genomic sequences used for the study are that of Human Pappillamovirus, which causes cervical cancer, among many other diseases.

Classification of Tabla Strokes Using Neural Network

The paper proposes classification of tabla strokes using multilayer feed forward artificial neural network. It uses 62 features extracted from the audio file as input units to the input layer and 13 tabla strokes as output units in the output layer. The classification has been done using dimension reduction and without using dimension reduction. The dimension reduction has been performed using Principal Component Analysis (PCA) which reduced the number of features from 62 to 28. The experiments have been performed on two sets of tabla strokes, which are played by professional tabla players, each comprises of 650 tabla strokes. The results demonstrate that correct classification of instances is more than 98 % in both the cases.

Classification and Rule-Based Approach to Diagnose Pulmonary Tuberculosis

The pulmonary tuberculosis (TB) is diagnosed conventionally from the test results obtained from different medical examinations. The paper proposes a novel methodology using the classification technique called Identification tree (IDT) to diagnose TB computationally. The model reduces the number of parameters required for the diagnosis substantially. It also offers a list of rules for the speedy and easy diagnosis. The effectiveness of the method has been validated by comparing with existing techniques using standard detection measures.

Genetic Algorithm to Generate Music Compositions: A Case Study with Tabla

The paper proposes a methodology to create valid music compositions using genetic algorithm. Indian percussion instrument tabla is used as a prototype for this purpose. For any given taal, the methodology could generate new compositions from randomly generated initial population of standard bols of the tabla. A unique alpha-numeric representation is used for string representation. Typical genetic operators like selection, crossover and mutation have been used, but with tailor made modifications to incorporate unique features of the instrument under study. Fitness function incorporates the concept of fuzzy string matching. Experiments were conducted using different taals and different population sizes. The computer-generated compositions have been validated by human experts for its validity and novelty.

Reduction of Risk Factors in Risk Assessment: An Identification Tree Approach

Abstract—The paper proposes a unique procedure to reduce risk factors in risk assessment. It offers a variant of decision tree called Identification tree for reducing number of risk factors used in assessment. The model, which uses auto insurance as a case study, employs historical evidences of different vehicles as risk factors. The work offers reduction of original risk factors from a set of twenty three to a reduced set of nine risk factors. The model was validated using real time and industry specific data.

Finding motifs using harmony search

The paper proposes a novel methodology for finding motifs of biological data. It uses music inspired meta-heuristic optimization technique called harmony search to find motif. The model is based on randomly generated l-mers as the initial harmony memory. Pitch adjustment and random selection are used to generate new l-mers, which are adjudged by a specially defined objective function. The proposed method is experimentally validated using sequences of Human Papillomavirus strains obtained from accredited and authorized sources.