Kalaavathi Institute of Technology B

Prof. Dr. Kalaavathi Institute of Technology B

orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
19 March 2025 in Neural Computing and Applications
Verified
Journal Article
Open access
PDF file
HKHEMANTH KARNATI
HEMANTH KARNATI
ASAnuraag Soma
Anuraag Soma
AAAdnan Alam
Adnan Alam
KBKalaavathi Institute of Technology B
Kalaavathi Institute of Technology B
Understanding and predicting air quality is pivotal for public health and environmental management, especially in urban areas like Delhi. This study utilizes a comprehensive dataset from the Central Pollution Control Board, detailing air pollutant concentrations at the ITO (Income Tax Office), Delhi station, from 2017 to 2023. Our focus on pollutants such as PM2.5, PM10, NO2, NH3, SO2, ozone, and CO highlights the challenges posed by missing data in environmental studies. Through the application of several imputation models, including KNN, linear regression, forward fill + backward fill, Fourier + KNN, linear interpolation, and statistical methods, we identified combination of forward fill and backward fill as the most effective method for addressing data gaps, specifically for ozone measurements. Building on this imputed dataset, we employ various forecasting models to predict PM2.5 levels, a critical pollutant. Our exploration includes time series and deep learning approaches, such as LSTM, Bi-LSTM, CNN-LSTM, GRU, Deep AR, WaveNet, TCN, ARIMA, SARIMA, and Prophet. The performance of these models is evaluated using daily data, with Bi-LSTM and LSTM with attention emerging as the top performers due to their accuracy in predicting PM2.5 concentrations. The implications of our findings extend beyond academic interest, offering practical insights for environmental policy and health advisories in Delhi. By demonstrating the efficacy of specific imputation and forecasting techniques, our research contributes to the broader effort of improving air quality monitoring and prediction. The success of Bi-LSTM and LSTM with attention models, in particular, suggests a promising avenue for future studies aiming to enhance the precision of environmental forecasts.


3 April 2023 in International Journal on Recent and Innovation Trends in Computing and Communication
Verified
Journal Article
Open access
PDF file
KBKalaavathi Institute of Technology B
Kalaavathi Institute of Technology B
BSB. Sridhevasenaathypathy
B. Sridhevasenaathypathy
NCNARENDER CHINTHAMU
NARENDER CHINTHAMU
DVDinesh Valluru
Dinesh Valluru
Sales is a basic standpoint for business growth. Demand for consumer products decides the success rate of every business resulting in a profit. Proper analysis of the consumer interest in a particular product decides future sales. The ordinary tactics for sales and promotion objectives no longer help businesses keep up with the speed of a challenging market because it goes out with no knowledge of consumer buying habits. As a consequence of technological developments, significant changes can be seen in the domains of marketing and selling. As a result of such developments, multiple important factors such as consumers' buying habits, target people, and forecasting sales for the coming years can be readily determined, assisting the sales crew in developing strategies to achieve an upsurge in their company. This paper investigates the use of Association Rule Learning with Feature Extraction to forecast sales performance in order to recognise buyers. The consumer's related goods are identified using the association framework. Data on buying activities are derived from purchase invoices provided by the business. The outcome of both is utilized to create a company strategy. Support, Confidence, and Lift are the metrics used for evaluating the quality of association rules produced by the model. Based on the buyers’ preferences this paper forecasts retail shop sales and predicts the association relation between the products by feature extraction with Association rule learning to improve future sales. The suggested approach is employed to discover the most common pairings of items found in the business. This will assist with promotion and revenue. This method can help you find intriguing cross-selling and connected goods. The WEKA tool was used to evaluate the correctness of the Association rule that was created.

27 October 2023 in International Journal on Recent and Innovation Trends in Computing and Communication
Verified
Journal Article
Open access
PDF file
KBKalaavathi Institute of Technology B
Kalaavathi Institute of Technology B
To avoid potential fatalities, it is vital to diagnose brain tumors at its earlier stage itself. Tumors vary widely in shape, size, and location and this makes accurate segmentation and classification difficult even with significant efforts in this field. Therefore, a hybrid method for classifying brain tumors called the hybrid VGG-XGBoost (hybrid VGG-XGB) model is proposed in this work. The K-means technique is used to extract significant characteristics from brain tumor images after they have first been improved using contrast-limited adaptive histogram equalization (CLAHE). Then, for extracting the high-level features segmented regions, VGG-19 model has been used and for classification process, XGBoost supervised learning technique has been utilized. As evidenced by criteria like precision, recall, F1-score, kappa score, and the confusion matrix, our novel approach outperformed baseline models and produced excellent results with a 98% accuracy rate. The suggested methodology for brain cancer diagnosis in healthcare systems is recommended by us, based on highly predictive outcomes

1 January 2026 in ITEGAM-Journal of Engineering and Technology for Industrial Applications (ITEGAM-JETIA)
Verified
Journal Article
Open access
PDF file
KBKalaavathi Institute of Technology B
Kalaavathi Institute of Technology B
MRM SujayKumar Reddy
M SujayKumar Reddy
AAArivoli A
Arivoli A
Understanding the behavior and emotions of dogs through their vocalizations is vital for various applications that includes emotional intelligence analysis and remote monitoring services. In this study, the vocal behavior of Dachshund dogs has been examined by using a novel feature extraction technique based on Mel Frequency Cepstrum Coefficients (MFCC) and advanced Deep Learning architectures. Classifying the vocalizations of pets accurately, particularly when they are left alone, is a critical challenge. Hence, an approach is offered in this paper that integrates MFCC-based feature extraction with multiple neural network architectures like CNN, LSTM, Feedforward Networks (FFN), and hybrid CNN- LSTM models. In comparison to the other architectures, the Stacked CNN-BiLSTM model outperforms better at detecting Pet Sound Events. This study enhances pet dog vocalization analysis, enabling better communication in the field.