Umangbhai Soni

Umangbhai Soni

Auto-generated profile, based on publicly available information

Umangbhai Soni

Faculty of Engineering and Technology, Parul Institute of Engineering and Technology, Parul University, Vadodara, India

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
26 December 2023 in Optical and Quantum Electronics
Verified
Journal Article
Closed access
PDF file
Abdulkarem AlamwganiAbdulkarem Alamwgani
Abdulkarem Alamwgani
Vishal SorathiyaVishal Sorathiya
Vishal Sorathiya
USUmangbhai Soni
Umangbhai Soni
AAAdam Alhawari
Adam Alhawari
MDMalek G. Daher
Malek G. Daher

1 December 2024 in Cybernetics and Information Technologies
Verified
Journal Article
Open access
PDF file
USUmangbhai Soni
Umangbhai Soni
GJGordhan Jethava
Gordhan Jethava
AGAmit P. Ganatra
Amit P. Ganatra
A loan is vital for individuals and organizations to meet their goals. However, financial institutions face challenges like managing losses and missed opportunities in loan decisions. A key issue is the imbalanced datasets in credit risk assessment, hindering accurate predictions of defaulters. Previous research has utilized machine learning techniques, including single or multiple classifier systems, ensemble methods, and class-balancing approaches. This review summarizes various factors and machine learning methods for assessing credit risk, presented in a tabular format to provide valuable insights for researchers. It covers data complexity, minority class distribution, sampling techniques, feature selection, and meta-learning parameters. The goal is to help develop novel algorithms that outperform existing methods. Even a slight improvement in defaulter prediction rates could significantly influence society by saving millions for lenders.

2 October 2023 in Plasmonics
Verified
Journal Article
Closed access
PDF file
Vishal SorathiyaVishal Sorathiya
Vishal Sorathiya
USUmangbhai Soni
Umangbhai Soni
VVVipul Vekariya
Vipul Vekariya
JGJaysheel Golani
Jaysheel Golani
Abdulkarem AlamwganiAbdulkarem Alamwgani
Abdulkarem Alamwgani
AAAdam Alhawari
Adam Alhawari

31 March 2026 in Advanced Engineering Letters
Verified
Journal Article
Open access
VDVipul Dave
Vipul Dave
PKPradeep Kumar Karsh
Pradeep Kumar Karsh
USUmangbhai Soni
Umangbhai Soni
VDVipul M Dabhi
Vipul M Dabhi
KZKhyati Zalawadia
Khyati Zalawadia
PRPatel Ankit Rameshbhai
Patel Ankit Rameshbhai
Time–frequency analysis is necessary for condition monitoring of rotating machinery; comparatively limited attention has been given to image- based texture feature analysis derived from time–frequency representations. A bearing fault diagnosis model is proposed that combines the Inverse Fast Walsh-Hadamard Transform (IFWHT), texture feature extraction, feature ranking, and machine learning classification. First, IFWHT transforms vibration signals into time-frequency images, thereby improving the effective representation of transient and fault- related data. Based on these images, two-dimensional texture attributes are derived from the Gray Level Co-occurrence Matrix (GLCM), which records the spatial relationships and structural patterns of various bearing health conditions. The ReliefF algorithm ranks and selects the most discriminative features, reducing feature redundancy and improving classification performance. Machine learning classifiers are then trained and evaluated using the ranked feature set. The findings show that the suggested ranked feature-based framework provides consistent and good-quality fault classification performance. Fine KNN and Cubic SVM are among the compared models, where the former will reach a maximum classification accuracy of 93.8%, and the latter will reach an accuracy of 91.7% at various combinations of faults. The methodology suggested offers a strong and adaptable approach in the context of fault diagnosis, which effectively promotes predictive maintenance by lessening unforeseen downtimes and increasing the dependability of operations in industries.


21 January 2026 in Discover Artificial Intelligence
Verified
Journal Article
Open access
PDF file
USUmangbhai Soni
Umangbhai Soni
GJGordhan Jethava
Gordhan Jethava
Credit risk assessment plays a crucial role in financial decision-making, but imbalanced datasets present significant challenges, leading to biased predictions and poor classification of defaulters. Traditional techniques such as under-sampling, over-sampling, and SMOTE aim to address this imbalance but suffer from issues like information loss, lack of diversity, and duplicated patterns. To address these issues, we propose a novel two-stage classification framework that integrates a Voting Classifier with Meta-Dynamic Ensemble Selection (Meta-DES). The first stage employs a voting classifier to make confident predictions while isolating ambiguous instances. The second stage applies Meta-DES to dynamically select the most competent classifiers based on meta-features and the region of competence. This approach enhances classification performance by refining uncertain predictions through meta-learning and leveraging ensemble techniques to improve robustness. Extensive experimentation on a real-world loan application dataset demonstrates that our approach significantly outperforms traditional machine learning models. Our method maintains reliable predictions as demonstrated by comprehensive performance evaluations. The proposed model achieves superior precision and recall, outperforming traditional classifiers. Experimental results highlight its balanced performance in distinguishing defaulters and non-defaulters, achieving an accuracy of 0.8332, with significantly improved F1 scores. Additionally, statistical validation using the McNemar test confirms the superiority of our approach over baseline models, with overwhelmingly significant results. By effectively reducing Type I and Type II errors, our approach enhances financial risk assessment and supports more reliable credit decision-making. The proposed methodology not only improves classification outcomes but also contributes to minimizing financial losses and optimizing lending strategies.