Jaya Thomas

Performance evaluation of SVM based semi-supervised classification algorithm

To construct decision boundaries for two-class classification, SVM approach is attractive due to its efficiency. However, this approach is useful for 2-class classification and when the classes (labels) for the data are known. In practice, we have collection of labeled as well as unlabelled data, and it gives rise to semi-supervised classification problem. In this paper, we give a semi-supervised classification algorithm based on support vector machine (SVM). Novel feature of our approach is the formulation of spherical decision boundaries and the exploitation of the dynamical system associated with support function to obtain the number of clusters. The experimental results on a few well-known datasets, namely, Iris dataset, Shuttle landing control dataset, Wisconsin Breast cancer dataset, glass dataset, and balance scale dataset, indicate that our approach results in satisfactory classification as well as generalization accuracy.

Overview of integrative analysis methods for heterogeneous data

In the big data era, data are not only generated in massive quantity but also in diversity. The heterogeneous characteristics of the diverse data sources on a subject provide complimentary information. However, they pose challenges in data analysis process. Then, what are the existing methods for utilizing theses heterogeneous data to improve data analysis and how can we choose amongst these methods? We categorize integrative methods for heterogeneous data analysis to Bayesian network based methods and multiple kernel based methods and describe them in detail with examples of successful applications in the bioinformatics field.

Design of efficient packing system using genetic algorithm based on hyper heuristic approach

In practice the maximum usage of container space arises in many applications which is one of the crucial economical requirements that have a wide impact on good transportation. A huge amount of monetary infrastructure is spent by companies on packing and transportation. This study recommends that there exists a scope for further optimization which if implemented can lead to huge saving. In this paper, we propose a new hyper heuristic approach which automates the design process for packing of two dimensional rectangular blocks. The paper contributes to the literature by introducing a new search technique where genetic algorithm is coupled with the hyper heuristic to get the optimal or sub optimal solution at an acceptable rate. The results obtained show the benefits of hyper-heuristic over traditional one when compared statistically on large benchmark dataset at the 5% level of significance. Improvements on the solution quality with high filling rate up to 99% are observed on benchmark instances.

Hybrid approach for 2d strip packing problem using genetic algorithm

In this paper we have studied the two-dimensional cutting stock problem, in which large number of small rectangles are to be placed in the big container such that the trim loss and height of the layout is minimized. We have proposed a placement approach along with a relevant fitness function to evaluate the overall goodness of the design layout. The computation results validate the solution and the effectiveness of the approach.

Solution to an attack on digital signature in SMS security

Short Message Service (SMS) is one of the mobile phone services widely used by the public worldwide. One of the main issues during the communication is information security and privacy. Digital Signature algorithms can be applied with the SMS cipher to prevent the message information from the repudiation attack. There are various existing algorithms for digital signature like RSA, DSA and ECDSA. Out of these, ECDSA has been proved and found the best algorithm in terms of efficiency and cryptanalysis. But ECDSA has some drawbacks also, and because of that some authors have proposed variant algorithms of ECDSA. In this paper, we propose solutions to an attack on digital signature that has been found on a variant of ECDSA. At the end of this paper, conclusion with the suitable solution approach with its security analysis is summarized.

Multi-kernel LS-SVM based integration bio-clinical data analysis and application to ovarian cancer

The medical research facilitates to acquire a diverse data types from the same individual for a particular cancer. Major challenge is how to integratively analyse the multiple data types. In this paper, we introduce a multiple kernel based pipeline for integrative analysis of four genomic data and a set of clinical data. In the pipeline, multiple-kernel is generated from the weighted sum of individual kernels and is used to stratify patients and predict clinical outcomes. We apply the pipeline on ovarian cancer data from TCGA and examine intra similarities of clinical factors of each subtype and calculate log-rank statistics to verify how well they cluster. We also examined the power of molecular and clinical data in predicting dichotomised overall survival data and tumour grade. It was observed that the integration of various data types yields better stratification and higher prediction accuracy as compared to using individual data types.

Maximizing information through multiple kernel-based heterogeneous data integration and applications to ovarian cancer

The medical research facilitates to acquire diverse type of data from the same individual for a particular cancer. Recent studies show that utilizing such diverse data results in more accurate predictions. The major challenge faced is how to utilize such diverse data sets in an effective way. In this paper, we introduce a multiple kernel based pipeline for integrative analysis of high-throughput molecular and clinical data. We apply the pipeline on Ovarian cancer data from TCGA. After multiple kernel have been generated from weighted sum of individual kernels, it is used to stratify patients and predict clinical outcomes. We examine the clinical outcomes of each subtype to verify how well they cluster.

Placement strategy for trim minimization in one-dimensional cutting stock

The cutting stock problem(CSP) is one of the famous problems of operational research. In the past few decades a number of techniques are proposed to solve the problem. Almost in all technique, the bottleneck is the column generation phase due to the enumerative nature. In this paper, we propose a technique, for column generation that reduces the total number of cutting pattern to be generated. The approach emphasizes the fact that the total number of pattern to be generated would be at most the number of stock requirement. In this approach we have used placement strategy i.e. taking each stock requirement at a time and checking relative position of other required stock. The results obtained are better than already existing delayed column generation, the trim loss have reduced for dataset varying from .001-.05%.

Semi-supervised classification for intrusion Detection System in networks

We propose a semi supervised classifier for intrusion detection. In our approach, we classify the data entering the computer network. To achieve this, we start with two broad classes of data namely, malicious data and good data. We use Support vector machine based classifier with spherical decision boundaries to classify a chosen subset of malicious data taken as training samples. In the Intrusion Detection System (IDS) database, all data identified as malicious data according to our classifier is included as signature (of attack). Using our classifier for testing the out-of-sample data samples, we observe that the accuracy of the system is 72% for web log data.

Impact of structural prior knowledge in SNV prediction: Towards causal variant finding in rare disease

Can structural information of proteins generate essential features for predicting the deleterious effect of a single nucleotide variant (SNV) independent of the known existence of the SNV in diseases? In this work, we answer the question by examining the performance of features generated from prior knowledge with the goal towards determining the pathogenic effect of rare variants in rare disease. We take the approach of prioritizing SNV loci focusing on protein structure-based features. The proposed structure-based features are generated from geometric, physical, chemical, and functional properties of the variant loci and structural neighbors of the loci utilizing multiple homologous structures. The performance of the structure-based features alone, trained on 80% of HumVar-HumDiv combination (HumVD-train) and tested on 20% of HumVar-HumDiv (HumVD-test), ClinVar and ClinVar rare variant rare disease (ClinVarRVRD) datasets, showed high levels of discernibility in determining the SNV’s pathogenic or benign effects on patients. Combined structure- and sequence-based features generated from prior knowledge on a random forest model further improved the F scores to 0.84 (HumVD-test), 0.75 (ClinVar), and 0.75 (ClinVarRVRD). Including features based on the difference between wild-type in addition to the features based on loci information increased the F score slightly more to 0.90 (HumVD-test), 0.78 (ClinVar), and 0.76 (ClinVarRVRD). The empirical examination and high F scores of the results based on loci information alone suggest that location of SNV plays a primary role in determining functional impact of mutation and that structure-based features can help enhance the prediction performance.