PSG College of Technology

About: PSG College of Technology is a based out in . It is known for research contribution in the topics: Machining & Thin film. The organization has 3174 authors who have published 3575 publications receiving 40690 citations.

Dye-sensitized solar cells with natural dyes extracted from rose petals

Abstract: Nanocrystalline TiO2 dye-sensitized solar cells have been fabricated using TiO2 photoelectrode sensitized using the extracts of red rose and table rose as natural sensitizers and their characteristics have been studied. The extracts having anthocyanin pigment (pelargonidin, peonidin and cyanidin), which have hydroxyl and carboxylic groups in the molecule can attach effectively to the surface of TiO2 film. The solar cell constructed using the red rose sensitized TiO2 photo-electrode exhibited a short-circuit photocurrent of 4.57 mA/cm2 and a power conversion efficiency of 0.81 % and that of table rose sensitized TiO2 photo-electrode exhibited a short-circuit photocurrent of 4.23 mA/cm2 and a power conversion efficiency of 0.67 %. Natural dye sensitized TiO2 photo electrodes present the prospect to be used as an environment-friendly, low-cost alternative system.

ECG arrhythmia classification based on logistic model tree

Abstract: This paper presents a diagnostic system for classification of cardiac arrhythmia from ECG data, using Logistic Model Tree (LMT) classifier. Clinically useful information in the ECG is found in the intervals and amplitudes of the characteristic waves. Any abnormality in the wave shape and duration of the wave features of the ECG is considered as arrhythmia. The ampli-tude and duration of the characteristic waves of the ECG can be more accurately obtained using Discrete Wavelet Transform (DWT) analysis. Further, the non-linear behavior of the cardiac system is well characterized by Heart Rate Variability (HRV). Hence, DWT and HRV techniques have been employed to extract a set of linear (time and frequency domain) and non-linear characteristic features from the ECG signals. These features are used as input to the LMT classifier to classify 11 different arrhyth-mias. The results obtained indicate an impressive prediction accuracy of 98%, validating the choice and combined use of the current popular techniques (DWT and HRV) for cardiac arrhythmia classification. The system can be de-ployed for practical use after validation by experts.

An enhanced ACO algorithm to select features for text categorization and its parallelization

Abstract: Feature selection is an indispensable preprocessing step for effective analysis of high dimensional data. It removes irrelevant features, improves the predictive accuracy and increases the comprehensibility of the model constructed by the classifiers sensitive to features. Finding an optimal feature subset for a problem in an outsized domain becomes intractable and many such feature selection problems have been shown to be NP-hard. Optimization algorithms are frequently designed for NP-hard problems to find nearly optimal solutions with a practical time complexity. This paper formulates the text feature selection problem as a combinatorial problem and proposes an Ant Colony Optimization (ACO) algorithm to find the nearly optimal solution for the same. It differs from the earlier algorithm by Aghdam et al. by including a heuristic function based on statistics and a local search. The algorithm aims at determining a solution that includes 'n' distinct features for each category. Optimization algorithms based on wrapper models show better results but the processes involved in them are time intensive. The availability of parallel architectures as a cluster of machines connected through fast Ethernet has increased the interest on parallelization of algorithms. The proposed ACO algorithm was parallelized and demonstrated with a cluster formed with a maximum of six machines. Documents from 20 newsgroup benchmark dataset were used for experimentation. Features selected by the proposed algorithm were evaluated using Naive bayes classifier and compared with the standard feature selection techniques. It was observed that the performance of the classifier had been improved with the features selected by the enhanced ACO and local search. Error of the classifier decreases over iterations and it was observed that the number of positive features increases with the number of iterations.