Dr. Premkumar Borugadda
@premkumar.b@srmap.edu.in
Post-Doctoral Fellow
SRM University-AP
I received my M.Tech. (CSE) and Ph.D. in the stream of Artificial Intelligence & Machine Learning in the Department of Computer Science at Pondicherry University, Karaikal Campus. Beside this, I also cleared the state eligibility test from both the states: AP (2012) and Tamilnadu (2012) and (2017)" in Computer Science and Applications. My publication includes 9 research papers in refereed journals and 19 research articles in national and international journals. Three book chapters have been accepted, and one book has been published with the title “Advanced Machine Learning Applications Using Python My research area is machine learning, deep learning, and computer vision. Research and teaching is my passions.
"Machine Learning Lectures for Beginners" This is my YouTube channel, where I explain various topics in machine learning and deep learning.
EDUCATION
M.Tech (CSE), Ph.D. AI & ML
RESEARCH, TEACHING, or OTHER INTERESTS
Artificial Intelligence, Computer Vision and Pattern Recognition, Agricultural and Biological Sciences, General Business, Management and Accounting
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Premkumar Borugadda, Ramasami Lakshmi, and Satyasangram Sahoo
Universiti Putra Malaysia
Tomato is the most popular and cultivated crop in the world. Nevertheless, the quality and quantity of tomato crops have been declining due to various diseases that afflict tomato crops. Hence, it becomes necessary to detect the disease early to prevent crop damage and increase the yield. The proposed model in this article predicts the infected tomato leaf images (9 classified diseases and also healthy class) obtained from the Plant Village dataset. In this model, Transfer learning was used to extract features from images by VGG16, yielding a high dimension of 25088 features. Overfitting is a commonly anticipated problem because of the higher dimensionality of data. To mitigate this problem, the authors have adopted a novel dimensional reduction-based technique: filter methods, feature extraction techniques like Principal Components Analysis (PCA), and the Boruta feature selection technique of wrapper methods. This adoption enables the proposed model to attain a significantly improved high accuracy of 95.68% and 95.79% in MLP and VGG16, respectively, by reducing its initial dimension on the tomato dataset containing 18160 images across 10 classes.
Snehit Shaprapawad, Premkumar Borugadda, and Nirmala Koshika
IEEE
In order to determine the worthiness of a car based on a variety of factors using machine learning models. In this study, the challenge is to prevent the model from becoming overfit and to generalize it. A combination of regularization techniques as well as hyperparameter tuning techniques was employed to overcome this challenge. Develop linear regression, lasso regression, ridge regression, elastic net regression, random forest, decision tree and Support Vector Machine models with hyperparameters. The objective of this article is to build a generalized model that can predict the price of used cars based on some factors, such as the car's mileage, the year it was made, the road tax, the type of fuel it uses, the size of its engine etc. Optimal model can help sellers, buyers, and car manufacturers. A relatively accurate prediction of price can be made based on information provided by users. Among the seven models, the support vector regressor is the optimal model based on the evaluation metrics such as R Squared (R^2) of 95.27%, Mean Absolute Error (MAE) of 0.142, Mean Squared Error (MSE) of 0.047, and Root Mean Squared Error (RMSE) of 0.218 at 90% of the train data and 10% of the validation data.
Sairam Vasa, PremKumar Borugadda, and Archana Koyyada
IEEE
A stroke is caused by a disturbance in blood flow to a specific location of the brain. This might occur due to an issue with the arteries. The objective of this research to develop the optimal model to predict brain stroke using Machine Learning Algorithms (MLA's), namely Logistic Regression (LR), Decision Tree Classifier (DTC), Random Forest Classifier (RFC), Support Vector Machine (SVC), Naive Bayes Classifier (NBC), KNN Classifier (KNN), and XGBoost Classifier (XGB).Apply the above algorithms with hyperparameter along with GridSearchCV (CV= 5) on the given dataset. The given dataset is imbalanced, while training the models, a few difficulties were met, including underfitting, a dataset with null values, and a model without balancing the data to boost performance of the models, need to balance the data by using a data sampling method such as SMOTE. Among the Seven models, XGB is the optimal model based on the accuracy of 96.34%.
Nirmala Koshiga, Premkumar Borugadda, and Snehit Shaprapawad
IEEE
The objective of this study is to create efficient machine learning (ML) models for the Heart Disease Prediction System (HDPS). This study shows how classification techniques for machine learning can forecast heart illness. To forecast and alert patients about potential cardiac abnormalities, machine learning (ML) models including Logistic Regression (LR), Decision Tree Classifiers (DTC), Random Forest Classifiers (RFC), Support Vector Classifiers (SVC), and voting classifiers are employed. Few challenges were encountered while developing the models, such as underfitting the model without balancing the data with decision tree classifier. The voting ensemble technique overcame the challenges and allowed for a generalized model on balanced data with high accuracy. The purpose of this investigation is to see whether the technique for properly forecasting heart disease is based on health factors. A voting classifier is made up of LR and RFC. Among all models, this voting classifier had the highest accuracy of 98.36%.
Satyasangram Sahoo, Prem Kumar B., and R. Lakshmi
IEEE
Transfer Learning by using Convolutional Neural Network has shown its outstanding performance in large scale image classification. India is a multi-script and multilingual country. Out of all languages, Telugu and Kannada have shared almost similar structure characters. Offline character recognition of both handwriting characters is a challenging task. Different feature extraction models have been used in character recognition studies. Convolutional Neural Network is used as efficiently supervised feature vector extraction. Feature vectors from large scale pre-trained ImageNet or COCO were proved more efficient than other script datasets for better result. Fine-tuning model of transfer learning was used in the studies for comparison studies.
Bichitrananda Behera, G. Kumaravelan, and Prem Kumar B.
IEEE
Document classification is a prevalent task in Natural Language Processing (NLP), which has an extensive range of applications in the biomedical domains such as biomedical literature indexing, automatic diagnosis codes assignment, tweets classification for public health topics, and patient safety reports classification. Nevertheless, manual classification of biomedical articles published every year into specific predefined categories becomes a cumbersome task. Hence, building an automatic document classification for biomedical databases emerges as a significant task among the scientific community. In recent years, Deep Learning (DL) models like Deep Neural Networks (DNN), Convolution Neural Networks (CNN), Recurrent Neural Networks (RNN), and Ensemble Deep Learning models are widely used in the area of text document classification for better classification performance compared to Machine Learning (ML) algorithms. The major advantage of using DL models in document classification is that it provides rich semantic and grammatical information for document representation through pre-trained word embedding. Hence, this paper investigates the deployment of the various state-of-the-art DL based classification models in automatic classification of benchmark biomedical datasets. Finally, the performance of all the aforementioned constitutional classifiers is compared and evaluated through the well-defined performance evaluation metrics such as accuracy, precision, recall, and f1measure.
Satyasangram Sahoo, , Prem Kumar Borugadda, Dr R Lakshmi, , and
Blue Eyes Intelligence Engineering and Sciences Engineering and Sciences Publication - BEIESP
The multi-input with mixed data modality of the model is based on three folded structure. The first input model is structured by Convolution Network that accepts the images related to the house. The implementation of the network is miniVGGNet design. The network is tested among, which gives a better outcome. The output valued is concatenated eventually with numerical value entry of the same set which is trained and processed by multi-layer perceptron for review the house price of the building. The linear activation is helped to evaluate the predicted value of price after equal dimension merging of convolutional and multi-layer perceptron network.