Sanjiban Sekhar Roy

Landscape Mapping of Functional Proteins in Insulin Signal Transduction and Insulin Resistance: A Network-Based Protein-Protein Interaction Analysis

The type 2 diabetes has increased rapidly in recent years throughout the world. The insulin signal transduction mechanism gets disrupted sometimes and its known as insulin-resistance. It is one of the primary causes associated with type-2 diabetes. The signaling mechanisms involved several proteins that include 7 major functional proteins such as INS, INSR, IRS1, IRS2, PIK3CA, Akt2, and GLUT4. Using these 7 principal proteins, multiple sequences alignment has been created. The scores between sequences also have been developed. We have constructed a phylogenetic tree and modified it with node and distance. Besides, we have generated sequence logos and ultimately developed the protein-protein interaction network. The small insulin signal transduction protein arrangement shows complex network between the functional proteins.

Stock Market Forecasting Using LASSO Linear Regression Model

Predicting stock exchange rates is receiving increasing attention and is a vital financial problem as it contributes to the development of effective strategies for stock exchange transactions. The forecasting of stock price movement in general is considered to be a thought-provoking and essential task for financial time series’ exploration. In this paper, a Least Absolute Shrinkage and Selection Operator (LASSO) method based on a linear regression model is proposed as a novel method to predict financial market behavior. LASSO method is able to produce sparse solutions and performs very well when the numbers of features are less as compared to the number of observations. Experiments were performed with Goldman Sachs Group Inc. stock to determine the efficiency of the model. The results indicate that the proposed model outperforms the ridge linear regression model.

Applicability of Rough Set Technique for Data Investigation and Optimization of Intrusion Detection System

The very idea of intrusion detection can be perceived through the hasty advancement following the expansion and revolution of artificial intelligence and soft computing. Thus, in order to analyze, detect, identify and hold up network attacks a network intrusion detection system based on rough set theory has been proposed in this article. In this paper we have shown how the rough set technique can be applied to reduce the redundancies in the dataset and optimize the Intrusion Detection System (IDS).

An effective hybridized classifier for breast cancer diagnosis

After lung cancer, breast cancer is known to be the greatest cause for death among females [20]. The improving effectiveness of machine learning approaches is being given a lot of importance by medical practitioners for breast cancer diagnosis. The paper proposes an effective hybridized classifier for breast cancer diagnosis. The classifier is made by combining an unsupervised artificial neural network (ANN) method named self organizing maps (SOM) with a supervised classifier called stochastic gradient descent (SGD). Also a comparative analysis is performed between the proposed approach and three supervised state of the art machine learning techniques decision tree (DTs), random forests (RF) and support vector machine (SVM). Initially SGD method is used in isolation for the classification task and then it is made to perform the classification after being hybridized with the unsupervised ANN technique on Wisconsin Breast Cancer Database (WBCD) [10]. The comparison is based up on classification accuracy that is produced by generating a confusion matrix. For verifying consistency of accuracy values, the classification task was repeated with Internet Advertisements Dataset [11]. The results of the classification experimentation using hybridization of SOM with SGD are much more superior to SGD in isolation. All the accuracy values have been computed after achieving a ten-fold cross validation on the both the datasets to further verify the classifiers performance.

Cancer data investigation using variable precision Rough set with flexible classification

The theory of Rough set is a new mathematical tool to deal with intelligent data mining proposed by Z Pawlak. This paper implements the concept of tolerance relation on incomplete information system using variable precision rough set(VPRS) with flexible classification, which is an extension to the classical rough set theory. Here we have analyzed a cancer data set provided by national cancer institute from (1975-2008) and estimated the value of tolerance factor for each country people based on their age, using VPRS with flexible classification model.

Classifying Spam Emails Using Artificial Intelligent Techniques

Spam emails have become an increasing difficulty for the entire web-users.These unsolicited messages waste the resources of network unnecessarily. Customarily, machine learning techniques are adopted for filtering email spam. This article examines the capabilities of the extreme learning machine (ELM) and support vector machine (SVM) for the classification of spam emails with the class level (d). The ELM method is an efficient model based on single layer feed-forward neural network, which can choose weights from hidden layers,randomly. Support vector machine is a strong statistical learning theory used frequently for classification. The performance of ELM has been compared with SVM. The comparative study examines accuracy, precision, recall, false positive, true positive.Moreover, a sensitivity analysis has been performed by ELM and SVM for spam email classification.

A Novel Diagnostic Approach Based on Support Vector Machine with Linear Kernel for Classifying the Erythemato-Squamous Disease

The diagnosis of the arythema disease is a real difficulty in dermatology. It causes redness induced in the lower level of the skin by hyperemia of the capillaries. It can harm several skin damages, inflammations. In this paper, we have put our efforts to design a diagnostic approach based on Support Vector Machine (SVM) with linear kernel by classifying the erythemato-squamous disease. SVM being a large margin classifier is a powerful pattern recognition and machine learning methodology that is widely used for both linear and non-linear classification problems. Comparing testing on different kernel methods, we have noticed that our method gives the better accuracy. Choosing the optimal value of the parameters is a crucial criterion and this was achieved by performing 3 fold cross-validations.

Can computational biology improve the phylogenetic analysis of insulin

Using computational biology, we have depicted the insulin phylogenetics. We have also analyzed the sequence alignment and sequence logos formation for both the insulin chain A and B for three groups namely, the mammalian group, vertebrates group and fish group. We have also analyzed cladograms of insulin for the mammalian group. In accordance with that path lengths, matrix for distance analysis, matching representation of nodes of the cladogram and dissimilarity between two nodes have been performed for both of the A and B chains of the mammalian group. Our results show that 12 amino acid residues (GlyA1, IleA2, ValA3, TyrA19, CysA20, AsnA21, LeuB6, GlyB8, LeuB11, ValB12, GlyB23 and PheB24) are highly conserved for all groups and among them some (GlyA1, IleA2, ValA3);(TyrA19, CysA20, AsnA21) are continuous. This study shows a rapid method to calculate the amino acid sequences in terms of evolutionary conservation rates as well as molecular phylogenetics.

A hypothetical relationship between the nuclear reprogramming factors for induced pluripotent stem (iPS) cells generation – bioinformatic and algorithmic approach

A hypothetical evolutionary relationship was generated between the nuclear reprogramming factors for induced pluripotent stem (iPS) cells generation. Utilizing bioinformatics techniques, sequence analyses and phylogenetic tree algorithms, a comparative study has been performed to understand the evolutionary relationship of human nuclear reprogramming factors of induced pluripotent stem cells (iPSCs) generation. Among the total six nuclear reprogramming factors, the four reprogramming factors (SOX2, C-MYC, KLF4, and LIN28) have significant evolutionary origin. Our study shows SOX2 and C-MYC have evolutionary relationship and common point of origin. Likewise, KLF4 and LIN28 are having evolutionary relationship and have common point of origin. Based on these evidences, we propose that our study may be a great help to the future researchers to understand the mechanism(s) as well as pathway of nuclear reprogramming process.

A Deep Learning Based Artificial Neural Network Approach for Intrusion Detection

Security of data is considered to be one of the most important concerns in today’s world. Data is vulnerable to various types of intrusion attacks that may reduce the utility of any network or systems. Constantly changing and the complicated nature of intrusion activities on computer networks cannot be dealt with IDSs that are currently operational. Identifying and preventing such attacks is one of the most challenging tasks. Deep Learning is one of the most effective machine learning techniques which is getting popular recently. This paper checks the potential capability of Deep Neural Network as a classifier for the different types of intrusion attacks. A comparative study has also been carried out with Support Vector Machine (SVM). The experimental results show that the accuracy of intrusion detection using Deep Neural Network is satisfactory.