Lipismita Panigrahi ¹, Kaberi Das ², Debahuti Mishra ³

Affiliations
1 Computer Science & Information Technology, Balasore College of Engineering and Technology, Balasore, Odisha, IN
2 Department of Computer Applications, Institute of Technical Education and Research, Siksha O Anusandhan Deemed to be University, Bhubaneswar, Odisha, IN
3 Dept. of Computer Sc. & Engineering, Institute of Technical Education and Research Siksha ‘O’ Anusandhan University, Bhubaneswar Odisha, IN

Abstract

Missing values can cause serious problems while mining data sets, such as i) loss of information and efficiency; ii) problem in data handling computation and analysis due to irregularities in the data patterns and non-applicability of standard software; and iii) serious bias if there are systematic differences between the observed and the unobserved data. Missing values can also cause misleading results by introducing bias. This paper focuses on a methodological framework for the development of an automated data imputation model based on Hybrid Higher Order Neural Network Classifier (HHONC). Four real, integer and simulated data sets are exposed to a perturbation experiment, based on the random generation of missing values. Here different imputation methods are applied in glass identification, wine recognition, heart disease and lung cancer data set to find the missing value and compared with different classic imputation procedures. The experiment not only improves the quality of a database with missing value but also the best results are clearly obtained with different variables.

Keywords

Hybrid Higher Order Neural Classifier (HHONC), Imputation Method, Missing Value, Neural Network(NN)

Full Text

   

Kaberi Das ¹, Jagannath Ray ¹, Debahuti Mishra ²

Affiliations
1 Computer Applications, Institute of Technical Education and Research, Siksha O Anusandhan University, Bhubaneswar, Odisha, IN
2 Computer Science & Engineering, Institute of Technical Education and Research, Siksha O Anusandhan University, Bhubaneswar, Odisha, IN

Abstract

This paper focuses on a methodological framework for gene selection by two approaches such as statistical approach and information based approach. Statistical measures are univariate measures where the gene relevance score of each gene is calculated without considering its co-relation (positive co-relation or negative co-relation) with other genes. Statistical approach includes Euclidian distance and Pearson co-relation. Mutual information is the measure of mutual dependence between two random variables in the case of probability theory. Information based approach includes information gain and dynamic relevance. In this paper the above gene selection methods are applied on four publicly available data sets such as, breast cancer, leukemia, hepatitis and dermatology to generate the subset of genes. Then, the resultant subset is fed through two classifiers namely Naive-Bayes and Support Vector Machine (SVM). Here also the data sets are directly applied to the classifier without applying the gene selection methods. Finally when we have compared the result, it has been found that all the data sets showing better accuracy when the classifiers are applied after gene selection technique which reflects the importance of gene selection technique.

Keywords

Information Based Approach, Naive Bayes, Statistical Approach, Support Vector Machine (SVM).

Full Text

   

Kailash Shaw ¹, Debahuti Mishra ¹

Affiliations
1 Computer Science & Engineering, Institute of Technical Education and Research, Siksha O Anusandhan University, Bhubaneswar - 751030, Odisha, IN

Abstract

Background/Objectives: Proteins are the fundamental units of biology; the mechanism by which primary sequence of proteins is predicted into its secondary structure is not yet accurately achieved. Methods/Statistical analysis: In this paper, BAT inspired FLANN (Functional Link Artificial Neural Network) model for protein secondary structure prediction with low computation cost and accuracy has been proposed. The proposed model consists of three different phases; i) First, the primary sequence of amino acid is converted into dynamic matrix for different window sizes then this dynamic matrix is used to derive correlation matrix, ii) Second, FLANN is used to classify each sequence of correlation matrix with different learning parameters and random weights. BAT inspired optimization algorithm has been used to optimize the weight and learning parameters of BAT-FLANN, and (iii) finally, refinement of secondary structure result. Results: Experiments were conducted with real datasets of some primary sequence on RS126 and CB396 datasets. Proposed method has been compared with existing DSC, NNSSP, PHD, PREDATOR, ZPRED, MULPRED, SVM models and found to be more promising. Conclusion/Application: The proposed method achieves average Q3 accuracy 81.2% and 82.7% for CB396 and RS126 dataset respectively. Moreover the segment overlap (SOV) is 76.1% and 75.3% for CB396 and RS126 dataset respectively.

Keywords

BAT-FLANN, Classifier, Dynamic Matrix, Protein Data Bank (PDB), Secondary Structure

Full Text

   

Subhadra Mishra ¹, Debahuti Mishra ¹, Gour Hari Santra ²

Affiliations
1 Siksha 'O' Anusandhan University, Bhubaneswar - 751030, Odisha, IN
2 Orissa University of Agriculture and Technology, Bhubaneswar - 751003, Odisha, IN

Abstract

Objective: This paper has been prepared as an effort to reassess the research studies on the relevance of machine learning techniques in the domain of agricultural crop production. Methods/Statistical Analysis: This method is a new approach for production of agricultural crop management. Accurate and timely forecasts of crop production are necessary for important policy decisions like import-export, pricing marketing distribution etc. which are issued by the directorate of economics and statistics. However one has understand that these prior estimates are not the objective estimates as these estimate requires lots of descriptive assessment based on many different qualitative factors. Hence there is a requirement to develop statistically sound objective prediction of crop production. That development in computing and information storage has provided large amount of data. Findings: The problem has been to intricate knowledge from this raw data , this has lead to the development of new approach and techniques such as machine learning that can be used to unite the knowledge of the data with crop yield evaluation. This research has been intended to evaluate these innovative techniques such that significant relationship can be found by their applications to the various variables present in the data base. Application / Improvement: The few techniques like artificial neural networks, Information Fuzzy Network, Decision Tree, Regression Analysis, Bayesian belief network. Time series analysis, Markov chain model, k-means clustering, k nearest neighbor, and support vector machine are applied in the domain of agriculture were presented.

Keywords

Artificial Neural Network, Decision Tree, Machine Learning, Regression Analysis, Time Series Analysis.

Full Text

   

I Vijaya ¹, Amiya Kumar Rath ², Bhagabat Puthal ³, Debahuti Mishra ¹, S. Satapathy ¹

Affiliations
1 Department of Computer Science and Engineering, Siksha ‘O’ Anusandhan University, Bhubaneswar - 751030, Odisha, IN
2 Department of Computer Science and Engineering and IT, V.S.S.U.T, Burla - 768018, Odisha, IN
3 Department of Computer Science and Engineering Retired Professor, IGIT, Sarang - 759146, Odisha, IN

Abstract

Objectives: A network is group of devices that are connected to each other called as nodes. The nodes can be mobile or static. The performance of mobile Adhoc wireless networks (MANETs) helps to identify the type of applications that are supported by the network. Our objective is Performance analysis of QoS parameters of MANETs on Mobility & Energy based Model with Routing Protocols. Method/Analysis: The various network scenarios of MANETS are simulated using NS2.35. Protocols used to analyze performance are AODV, DSDV and DSR. Network layer parameters (throughput, packet delivery ratio, normalized routing overhead and average end-to-end delay) are evaluated. Network scenarios are generated through variation in pause time and number of nodes. Area of simulation is formed in 600*600 m*m area. Findings/ Results: The mobiles devices in the network get connected only when there is a demand for it. The reactive gateway discovery algorithm is used in AODV and DSR. With the random movement of nodes in the simulated area (direction) and variation in mobility, the delay and packet drop increases but PDR and throughput decreases. There is a significant differential observed while measuring the performance. Our observation with respect to DSR was it reacted well for two parameters delivery ratio and routing overhead. Average delay was less in AODV and DSDV performed well providing loop free path. Conclusion: After the simulation study and all experimental evaluations we can conclude that the DSR protocol dominates all other protocols like AODV and DSDV. The Dynamic Source Routing protocol in mobility and energy based model for throughput, packet delivery ratio performs well than AODV and DSDV. The adverse result is with the increase of node speed, routing overhead increased for DSR. Positive aspect of DSR was that average energy consumption was quite low in contrast to AODV and DSDV.

Keywords

Average Energy Consumption, Average End-To-End Delay, MANET, Normalized Routing Overhead, Packet Delivery Ratio, Throughput.

Full Text

   

Smita Prava Mishra ¹, Debahuti Mishra ², Srikanta Patnaik ²

Affiliations
1 Computer Science and Information Technology, Institute of Technical Education and Research, Siksha O Anusandhan University, Bhubaneswar – 751003, Odisha, IN
2 Computer Science and Engineering, Institute of Technical Education and Research, Siksha O Anusandhan University, Bhubaneswar – 751003, Odisha, IN

Abstract

The researchers of the data mining domain presume that the study of traditional clustering techniques is saturating day by day. But, a deep insight into those techniques unfolds many silhouettes which could lead to many more applications in diverged domains. In clustering, the attributes of the data provide the information needed for data segregation. There may exist some real world data with less number of attributes but more information contained in them and may be of interest for some applications. Because of less number of attributes, the data may not be well separated by any of the clustering techniques. Data expansion techniques are methods for constructing more number of attributes from less number of attributes. With the application of these techniques, an expanded data set may be reconstructed from a given data set during data preprocessing. The current work pronounces the fact that, the expanded data at times yield better clustering results than the real data. This paper is an attempt to empirically evaluate and analyze the effects of data expansion on clustering results where validity of the results are established through internal indexing techniques and probabilistic validation measures.

Keywords

Cluster Analysis, Cluster Validity, Data Expansion, Internal Indexing, Probabilistic Measures

Full Text

   

Pradeep Kumar Mallick ¹, Debahuti Mishra ¹, Srikanta Patnaik ¹, Kailash Shaw ¹

Affiliations
1 Department of Computer Science and Engineering, Siksha ‘O’ Anusandhan University, Bhubaneswar - 751030, Odisha, IN

Abstract

Objectives: This study proposes a hybrid model of simultaneous gene clustering and gene selection for gene expression datasets using hierarchical clustering and rough set theory for classification of data patterns. Methods/Analysis: The internal architecture of the proposed model broadly works in three phases, in first phase; the initial clusters are formed using hierarchical clustering and again those resulted clusters are divided into more clusters using based on lower and upper approximation property of rough set theory. In second phase; the reduct property of rough set is applied on obtained clusters from the second phase; and in third phase, the gene ranking and cluster ranking has been employed to rank the genes in clusters to discover significant of informative genes. This method tries to find the genes of interest known as significant genes and maximize the accuracy of the model with reduction percentage. The advantage of this approach is analyzed by experimental results on two benchmark datasets such as Leukemia and Colon Cancer. Finally, the classification performance of the original datasets were recorded using Support Vector Machine (SVM) classifier and also with few existing feature/gene selection and clustering techniques. Findings: The experimental results and performance measures proves the efficiency of the proposed hybridized technique over existing feature/gene selection as well as established traditional k-means clustering technique.

Keywords

Gene Selection, Hierarchical Clustering, Lower Approximation, Reduct, Rough Set Theory, Upper Approximation.

Full Text

   

Ramamani Tripathy ¹, Debahuti Mishra ², V. Badireenath Konkimalla ³

Affiliations
1 Department of Computer Science and Engineering, Siksha O Anusandhan University, Bhubaneswar, Odisha, IN
2 National Institute of Science Education and Research (NISER), Jatni - 752050, Odisha, IN
3 Department of Atomic Energy, Bhubaneswar, Odisha, IN

Abstract

Objectives: The current study is focused on design of a computational model for human ABC transporters; wherein the TM-sequences matching the CRAC/CARC motif are extracted. Methods: The postulation of cholesterol binding motif (CRAC/CARC), its presence in different proteins and validating its interaction with cholesterol has indeed established the importance of the motif in cholesterol-mediated modulation of protein/signaling pathway. Several viral proteins and membrane proteins (especially alpha-helical trans membrane proteins) such as GPCR transporters are reported to be modulated by cholesterol. The experimental studies are so far performed on only a few proteins in a family but based on an evolutionary conservation and consensus an exploration can be done confidently within a family. However, the representation of motif has a low consensus yielding several false positives thus reducing its reliability. Findings: A computational hybrid clustering method based on rough set with fuzzy c-means algorithm is used to mine the cholesterol sequence from ABC family. Higher weightage is given to those sequences based on the following parameters: motifs with more number of sub motifs, number of helices bearing the motif in a protein and compliance with the orientation of the cholesterol in the membrane for its interaction with the motif. Improvement: A detailed study in a given super family with an approach to reduce redundancy and enrichment can improve its predictability.

Keywords

ABC transporter, CRAC/CARC, Fuzzy c-Means, GPCR, Motif, Rough Set.

Full Text