Amiya Kumar Rath

A Naïve SVM-KNN based stock market trend reversal analysis for Indian benchmark indices

This paper proposes a hybridized framework of Support Vector Machine (SVM) with K-Nearest Neighbor approach for Indian stock market indices prediction. The objective of this paper is to get in-depth knowledge in the stock market in Indian Scenario with the two indices such as, Bombay Stock Exchange (BSE Sensex) and CNX Nifty using technical analysis methods and tools such as predicting closing price, volatility and momentum of the stock market for the available data. This hybrid model uses SVM with different kernel functions to predict profit or loss, and the output of SVM helps to compute best nearest neighbor from the training set to predict future of stock value in the horizon of 1 day, 1 week and 1 month. The proposed SVM and KNN based prediction model is experienced with the above mentioned distinguished stock market indices and the performance of proposed model has been computed using Mean Squared Error and also been compared with recent developed models such as FLIT2NS and CEFLANN respectively. The limitation of both of those existing models undergoes complex weight updating procedures, whereas, proposed SVM-KNN hybridized model scales relatively well to high dimensional data and the trade-off between classifier complexity and error can be controlled explicitly and have better prediction capability.

Comparative analysis of SpatialHadoop and GeoSpark for geospatial big data analytics

In this digitalised world where every information is stored, the data a are growing exponentially. It is estimated that data are doubles itself every two years. Geospatial data are one of the prime contributors to the big data scenario. There are numerous tools of the big data analytics. But not all the big data analytics tools are capabilities to handle geospatial big data. In the present paper, it has been discussed about the recent two popular open source geospatial big data analytical tools i.e. SpatialHadoop and GeoSpark which can be used for analysis and process the geospatial big data in efficient manner. It has compared the architectural view of SpatialHadoop and GeoSpark. Through the architectural comparison, it has also summarised the merits and demerits of these tools according the execution times and volume of the data which has been used.

Far efficient K-means clustering algorithm

Clustering in data analysis means data with similar features are grouped together within a particular valid cluster. Each cluster consists of data that are more similar among themselves and dissimilar to data of other clusters. Clustering can be viewed as an unsupervised learning concept from machine learning perspective. In this paper, we have proposed an effective method to obtain better clustering with much reduced complexity. We have evaluated the performances of the classical K-Means approach of data clustering and the proposed Far Efficient K-Means method. The accuracy of both these algorithms were examined taking several data sets taken from UCI [13] repository of machine learning databases. Their clustering efficiency has been compared in conjunction with two typical cluster validity indices, namely the Davies-Bouldin Index and the Dunns Index for different number of clusters, and our experimental results demonstrated that the quality of clustering by proposed method is much efficient than K-Means algorithm when larger data sets with more number of attributes are taken into consideration.

Simulation and performance evaluation of AODV, DSDV and DSR in TCP and UDP environment

An ad hoc network is a collection of wireless mobile nodes dynamically forming a temporary network without the use of any existing network infrastructure or centralized administration. A number of routing protocols such as Dynamic Source Routing (DSR), Ad Hoc on-Demand Distance Vector Routing (AODV) and Destination-Sequenced Distance-Vector (DSDV) have been implemented. In this paper, an attempt has been made to compare the performance of two prominent on-demand reactive routing protocols: DSR and AODV, and proactive DSDV protocol in TCP (Transmission Control Protocol) and UDP (User Datagram Protocol) environments. A simulation model with Media Access Control (MAC) is used to study interlayer interactions and their performance implications. These include the measurement of network throughput using TCP and UDP as well as the delay in transmission of packets using TCP. Correlation between the two sets of results is found to be satisfactory enough to validate the simulation process. Given this validation, based on similar simulation techniques, the investigation of a larger scale Ad-hoc network is then carried out. These simulations are carried out using NS-2 simulator. The results presented in this work illustrate the importance in carefully evaluating and implementing routing protocols in an ad hoc environment.

Assessment of basic clustering techniques using teaching-learning-based optimisation

There has been lot of talk regarding the initial cluster centre selection, because a bad centroid may result in malicious clustering. Due to this reason, we have taken the help of a latest population-based evolutionary optimisation technique called teaching-learning-based optimisation TLBO for selecting near about optimum cluster centres. After getting the finest initial centroids, we perform the necessary clustering by means of our proposed Enhanced clustering algorithm. In this paper, we have evaluated and assessed the performances of three different TLBO-based clustering algorithms: TLBO-supported classical K-means, TLBO-based fuzzy c-mean and our proposed approach of TLBO-based data clustering. Their clustering efficiency has been compared in conjunction with two typical cluster validity indices, namely the Davies-Bouldins index and the Dunns index. We extend our comparison by taking into account their calculated average quantisation error. Each algorithm is then tested on several datasets taken from UCI repository of machine learning databases. Experimental results show that our proposed approach produces better clustering with minimum quantisation error for most of the datasets as compared to the other discussed methods. Also the problem of initial centre selection is minimised to a greater extent.

Algorithm aspects of dynamic coordination of beacons in localization of Wireless Sensor Networks

Wireless Sensor Networks(WSNs) consists of hundreds of nodes which are of low power, low cost, and tiny devices. The main functionality of these nodes is to sense the environment and send the sensed data to the observer. In order to validate and get the significance of sensing data, location information of the sensor node needs to be combined with the sensed data. In addition to this, there are many other issues of WSN such as routing, coverage, etc. which also need the location information of sensor nodes. Several approaches, including range-based and range-free, have been proposed to calculate positions for randomly deployed sensor nodes. In this paper, we proposed a distributed technique for localization of sensor nodes using few mobile anchor nodes. These mobile anchor nodes move in the network space and periodically broadcast beacon messages about their location. Static sensor nodes receive these messages as soon as they come under the communication range of any mobile anchor node and compute their position based on the range based technique. Another contribution of this paper is to identify the importance of mobile anchor node over static anchor node in localization. The performance of the proposed algorithm is carried out using the Castalia simulator. The simulation result shows that mobile anchor node provide better accuracy as compared to static anchor node for sensor node localization.

Self-assessment Model and Review Technique for SPICE: SMART SPICE

Self-assessment to understand the capability of any organization or “Where we Stand” is the key objective of this paper. It enhances the software engineering practices in the context of the quality of products. Software Process Improvement Assessment frameworks like CMMI and ISO/IEC 15504 –SPICE does assessment by the experts and professional involving high cost resulting many SMEs refrain from adapting the standard models for assessment. Perceived need observed to have a Self-assessment model, which will enable the organization using indigenous model for assessment following standards. The proposed model will address the issues of strength and weak-ness and to take appropriate measure for software process improvement thus enhancing organization’s capability. It is a well disciplined and well targeted assessment framework based on basic and simple questionnaires that covers the whole organization diversified activities and success stories to arrive at conclusion and establishing improvement initiatives.

CPB: A Model for Biclustering

Mining biclusters that exhibit both consistent trends and trends with similar degrees of fluctuations is vital to bioinformatics research. However, existing biclustering methods are not very efficient and effective at mining such biclusters. Most biclustering models, including those used in subspace clustering, define similarity among different objects by distances over either all or only a subset of dimensions in gene expression data. However, distance functions are not always adequate in capturing co-relations among the objects. In fact, strong co-relations may still exist among a set of objects even if they are far apart from each other as measured by the distance function.Under the CPB (Coherent Pattern Biclustering) model, we proposed, two objects are similar if they exhibit coherent pattern on a subset of dimensions. For instances, in DNA microarray analysis, the expression levels of two genes may rise or fall synchronously in response to a set of environmental stimuli. Though the magnitude of their expression levels may not be close, but the pattern they exhibit can be very much similar. Our proposed model is interested in finding such coherent patterns of biclusters of genes and with a general understanding of biological processes that many genes participate in multiple different processes.

Feature reduction using principal component analysis for agricultural data set

Many applications like video surveillance, telecommunication, weather forecasting and sensor networks uses high volume of data of different types. The effective and efficient analysis of data in such different forms becomes a challenging task. Analysis of such large expression data gives rise to a number of new computational challenges not only due to the increase in number of data objects but also due to the increase in number of attributes. Hence, to improve the efficiency and accuracy of mining task on high dimensional data, the data must be preprocessed by an efficient dimensionality reduction method. In this paper, we have proposed to use the method of k-means clustering and principal component analysis (PCA) approach for attribute reduction, which initially apply PCA to obtain reduced uncorrelated attributes specifying maximal eigenvalues in the dataset with minimum loss of information. Then again we proposed to use k-means on the PCA reduced dataset to discover discriminative features that will be the most adequate ones for classification. This is a combination of clustering approach with feature reduction to obtain a minimal set attributes retaining a suitably high accuracy in representing the original features. We have used the Greengram agricultural data set. Finally, we found that the result of clustering is same after reducing the attributes using PCA.

Generating PNS for Secret Key Cryptography Using Cellular Automaton

The paper presents new results concerning application of cellular automata (CAs) to the secret key using vernam cipher cryptography.CA are applied to generate pseudo-random numbers sequence (PNS) which is used during the encryption process. One dimensional, non-uniform CAs is considered as a generator of pseudorandom number sequences (PNSs) used in cryptography with the secret key. The quality of PNSs highly depends on a set of applied CA rules. Rules of radius r = 1 and 2 for non-uniform one dimensional CAs have been considered. The search of rules is performed with use of evolutionary technique called cellular programming. As the result of collective behavior of discovered set of CA rules very high quality PNSs are generated. The quality of PNSs outperforms the quality of known one dimensional CA-based PNS generators used in the secret key cryptography. The extended set of CA rules which was found makes the cryptography system much more resistant on breaking a cryptography key.