Geeta Sikka

Recent Techniques of Clustering of Time Series Data: A Survey

Time-Series clustering is one of the important concepts of data mining that is used to gain insight into the mechanism that generate the time-series and predicting the future values of the given time-series. Time-series data are frequently very large and elements of these kinds of data have temporal ordering. The clustering of time series is organized into three groups depending upon whether they work directly on raw data either in frequency or time domain, indirectly with the features extracted from the raw data or with model built from raw data. In this paper, we have shown the survey and summarization of previous work that investigated the clustering of time series in various application domains ranging from science, engineering, business, finance, economic, health care, to government.

Estimating function points: Using machine learning and regression models

Function Point Analysis (FPA) is one of the most reliable methods for measuring the size of computer software. It is extensively being used as Industry standard for sizing. It is also tremendously useful in estimating projects, managing change in requirements, measuring efficiency and communicating functional requirements. International Standard bodies like International Function Point User Group (IFPUG) has been maintaining a repository of data based projects with measures like data functions, transactional functions, function size, work effort, delivery date, project duration, max team size, development technique and platform etc. This paper presents the results of analysis that were performed on IFPUG data using a Statistical Package. Modelling techniques like Multivariate Adaptive Regression Splines (MAR Splines), Support Vector Machine (SVM), Automated Neural Network (ANN), k-Nearest Neighborhood (kNN) were used to estimate the function points from the IFPUG data set repository The training and validation data was randomly selected from the data repository. The performance of the various models was analyzed by comparing the observed and predicted function points. The comparison of results on the validation data set of 100 projects indicate that MAR Splines gives the best predicted values for function points as it has the lowest Mean Relative Error (MRE) and highest correlation coefficient. SVM and ANN also gave good results as compared to kNN. MAR Splines is thus a competitive predictive model for estimating function points. Thus, these machine learning and regression models can be used as approximation tool for function point metric.

An image interpolation based reversible data hiding scheme using pixel value adjusting feature

In this paper, we propose an image interpolation based reversible data hiding scheme using pixel value adjusting feature. This scheme consists of two phases, namely: image interpolation and data hiding. In order to interpolate the original image, we propose a new image interpolation method which is based on the existing neighbor mean interpolation method. Our interpolation method takes into account all the neighboring pixels like the NMI method. However, it uses different weight-age as per their proximity. Thus, it provides the better quality interpolated image. In case of data hiding phase, secret data is embedded in the interpolated pixels in two passes. In the first pass, it embeds the secret data into the odd valued pixels and then in the second pass, the even valued pixels are used to embed the secret data. To ensure the reversibility of the proposed scheme, the location map is constructed for every pass. Basically, the proposed scheme only increases/decreases the pixel values during data hiding phase, which improves the performance of the proposed scheme in terms of computation complexity. Experimentally, our scheme is superior to the existing scheme in terms of data hiding capacity, image quality and computation complexity.

Opinion mining of news headlines using SentiWordNet

Opinion Mining (also known as “Sentiment Analysis”) is an area of text classification which continuously gives its contribution in research field. The main objective of Opinion mining is Sentiment Classification i.e. to classify the opinion into positive or negative classes. SentiWordNet is an opinion lexicon derived from the WordNet database where each term is associated with some numerical scores indicating positive and negative sentiment information. Up until recently most researchers presented opinion mining of online user generated data like reviews, blogs, comments, articles etc. Opinion mining for offline user generated data like newspaper is unconcerned so far despite the fact that it is also explored by many users. As a first step, this paper present opinion mining for newspaper headlines using SentiWordNet. Further, most of the researchers implement the opinion mining by separating out the adverb-adjective combination present in the statements or classifying the verbs of statements. On the other hand, in this paper we analyze each and every word in the News headline whether it is a noun, verb, adverb, adjective or any other part-of-speech. During experiment, python packages are used to classify words. Then SentiWordNet 3.0 is used to identify the positive and negative score of each word thus evaluating the total positive/negative impact in that news headline.

Recent methods for software effort estimation by analogy

Fairly accurate cost and effort predictions of software projects have always been a challenging goal for both, industry as well as academia. Most approaches for effort estimation are either algorithmic or analogy based. The most well-known algorithmic models are COCOMO 81 [1] and Function Points [2]. Estimation by analogy, on the other hand, is essentially a form of case based reasoning. Fuzzy logic, Grey System Theory, Machine Learning techniques such as Genetic Algorithms, Support vector Machines, etc. have been used to optimize the prediction by analogy method. We study these analogy based approaches to software effort estimation and compare some of these techniques on the basis of widely used measures of accuracy.

A high payload data hiding scheme based on modified AMBTC technique

In this paper, we propose a data hiding scheme which uses our modified AMBTC compression technique for embedding the secret data. Our modified AMBTC technique converts the one bit plane into two bit plane which helps in achieving better quality compressed image as well as high capacity. In this scheme, we first apply the original AMBTC technique on the given cover image then identify the smooth and complex blocks using a user defined threshold value. In case of the smooth blocks, it converts the one bit plane into two bit plane using mean value of the block and replaces all the bits of the bit plane with the secret data bits. It calculates four quantization levels in place of two old quantization levels. In case of complex blocks, it converts the one bit plane into two bit plane but here only the first LSBs of the newly constructed bit plane is replaced by the secret data bits. The four new quantization levels are calculated using the resultant bit plane. Thus, this scheme is able to embed 2 bits into each pixel of the smooth blocks and one bit in each pixel of complex blocks. It provides good quality stego image because the introduced error during the secret data embedding is reduced by having four quantization levels. Experimentally, our scheme is superior to the existing AMBTC based data hiding schemes in terms of both data hiding capacity and image quality. In fact, the proposed scheme hides approximately two times more secret data than the existing schemes with better image quality.

A survey on wireless sensor network clustering protocols optimized via game theory

Wireless sensor network (WSN) consists of low size, power constrained nodes that sense the environment and communicate this information through wireless links. There are a number of research issues in WSN with energy efficiency being one of the prime issues for WSN applications. In clustering-based routing protocols, cluster head selection has significant effect on performance of the protocol, along with routing technique. Game theory as a mathematical notion, being able to analyze interactive decision situations, is applicable to a wide spectrum within WSN. It can assist in designing more efficient protocols. This article surveys the application of game theory in wireless sensor network protocols, specifically in the domain of communication protocols involving cluster formation i.e. clustering protocols in WSN and how it optimizes the functioning of these protocols.

A Study on Vectorization Methods for Multicore SIMD Architecture Provided by Compilers

SIMD vectorization has received important attention within the last few years as a vital technique to accelerate multimedia, scientific applications and embedded applications on SIMD architectures. SIMD has extensive applications; though the majority and focus has been on multimedia. As a result of it is an area of computing that desires the maximum amount of computing power as possible, and in most of the cases, it is necessary to compute plenty of data at one go. This makes it an honest candidate for parallelization. There are many compiler frameworks which allow vectorization such as Intel ICC, GNU GCC and LLVM etc. In this paper, we will discuss about GNU GCC and LLVM compilers, optimization methods, vectorization methods and evaluate the impact of various vectorization methods supported by these compilers and at last note we will discuss about the methods to enhance the vectorization process.

An AMBTC compression based data hiding scheme using pixel value adjusting strategy

Image steganography is one of the most important research areas of information security where secret data is embedded in the images to conceal its existence while getting the minimum possible statistical detectability. To achieve a good tradeoff between the hiding capacity and image quality, more work needs to be further researched. In this paper, we propose high capacity data hiding scheme by employing the absolute moment block truncation coding (AMBTC) compression. It exploits the redundancy of blocks of the AMBTC-compressed image to embed the secret data. The pixel values of the AMBTC-compressed image are modified at most by one for hiding the secret data. Thus, it is able to maintain the stego-image quality after hiding the secret data. Experimental results validate the effectiveness of the proposed scheme and show that it outperforms various existing methods in terms of both hiding capacity and stego-image quality.

Image interpolation based high capacity reversible data hiding scheme

In this paper, we propose a new interpolation technique which considers all the neighboring pixels as well as their impact on the reference pixels to provide better quality interpolated image and a new data hiding scheme which embeds the secret data in the interpolated pixels by taking into account the human visual system so that quality of the resultant image is maintained. The proposed interpolation technique is an improvement of the existing neighbor mean interpolation (NMI) technique in such a way that the interpolated image would have more resemblance to the input image. The proposed interpolation technique has less computational cost like NMI as it does not perform much computation during estimation unlike B-Spline, Bilinear Interpolation etc. The proposed data hiding scheme comes into the category of reversible data hiding scheme as the input image can be reconstructed after extraction of the entire secret data at the receiver side. Thus, it reduces the communication cost. Furthermore, the proposed data hiding scheme identifies the smooth and complex regions of the interpolated (or cover) image by dividing the same into blocks. It then embeds more bits into the complex regions of the image so that data hiding capacity as well as the image quality can be enhanced. The experimental results shows that the percentage increment in the PSNR value and capacity of the proposed scheme with respect to Chang et al. method is in the range of 0.26 to 30.60% and 0.87 to 73.82%, respectively. Moreover, the modified NMI yields higher PSNRs than other interpolating methods such as NMI, BI, and ENMI.