Manohar Mishra

Financial markets are worldwide interconnected, and trends of one market often influence others through cross-country market movements. Financial Markets Forecasting (FMF) models that ignore these dependencies risk may lose predictive accuracy. In this study, a novel framework that combines Wavelet Δ-band decomposition, Light Gradient Boosting Machine (LGBM), and Constricted Particle Swarm Optimization (CPSO) is presented to forecast stock indices while explicitly modelling cross-country effects. In this regard, the approach decomposes the target series into multi-resolution bands using discrete wavelet transform (DWT), enabling noise reduction and separation of frequency dependent trends. Afterward, each band is modelled with CPSO-tuned LGBM regressors, and predictions are aggregated via inverse DWT. Notably, the framework incorporates US market signals (Dow Jones Index) along with Indian market features to predict the closing price of three different indices of national stock exchange of India (NSEI), capturing global-to-local dependencies. Experimental evaluation on more than of ten years of data with 18 technical indicators shows that the proposed DWT–CPSO–LGBM approach achieves strong predictive performance (for instance of Nifty-50 dataset: RMSE ≈82.46, MAE ≈57.31, and MAPE <0.23% on the last 20 test samples). The ablation study compares the wavelet-free, hyperparameter tuning free and baseline models, where the proposed method demonstrates clear improvements. The findings of this work clearly highlight both methodological novelty and practical value in developing forecasting tools that incorporate cross-country data for stock market investors.

In 1997, Long Short-Term Memory (LSTM) networks were proposed, which significantly changed the landscape of sequential data analysis by resolving the critical issue of the vanishing gradient problem in recurrent neural networks (RNNs). Over the last 25 years, LSTM has advanced from its inception as an innovative solution to its widespread adoption as an essential tool in various fields, including natural language processing (NLP), speech recognition, financial prediction, and healthcare analytics. The present study is a bibliometric review of the evolution of LSTMs. The evolution of LSTM is discussed in terms of its theoretical advancements, architectural developments, and its applications. The study is based on data obtained from the Scopus database, which is then analyzed to identify publication patterns, prominent authors, prominent institutions, and global contributions to the field. The present study is an insightful review of the evolution of LSTM, highlighting its developments and advancements, as well as its applications, to identify its future scope.

This study presents a novel algorithm for automatic fault detection in multi-terminal voltage source converter-based high voltage direct current (MT-VSCHVDC) systems. The approach integrates the wavelet scattering transform (WST) to extract low-variance feature vectors and a newly developed variable batch size long-short-term-memory (VB-LSTM) network for accurate fault detection. The synthetic minority oversampling technique addresses class imbalance issues in fault and no-fault data. Utilizing a 5-fold cross-validation process, the study demonstrates that WST-based features, combined with the proposed VB-LSTM network and a hybrid training strategy, achieve 100% and 99.46% accuracy in classifying internal and external faults with no-noise and 20db noisy conditions respectively. The variable epoch size training enhances convergence speed, leading to more stable and consistent results. A secondary LSTM model with a similar layer architecture is also trained and evaluated on WST-based features to identify the fault location within internal faults. The proposed fault distance (WST-VB-LSTM) estimation model achieves excellent performance with an R-square of 0.9946 and MAE of 0.0135 without noise, and maintains robustness under noise with 30 dB (R² = 0.9903, MAE = 0.0182) and 20 dB (R² = 0.9598, MAE = 0.0364). Lastly, comparative performance analysis reveals that the proposed model outperforms recently published works and state-of-the-art techniques, exhibiting higher accuracy and significantly lower error metrics across varying noise conditions.

In this paper, a non-unit protection approach is suggested for recognizing fault events (internal and external) and their location in a High Voltage Direct Current (HVDC) system employing a deep learning (DL) model. This work utilizes a small sample size of DC current signals from one end of an HVDC-fed system, achieving a balance between data manageability and preserving vital information related to faults. The simulation of fault samples is carried out with the maximum possible operating scenarios, such as different fault resistances, locations, types of faults, and noisy conditions. Afterward, combined data with varying noise values is trained by the proposed DL model. The testing is performed under individual noise levels to evaluate model robustness and accuracy in real-time scenarios. In the DL framework, the work focuses on two core architectures, Convolutional Neural Networks (CNN) and Transformer models, and assesses their performance individually. Furthermore, it explores the hybrid model CNN-Transformer with a multi-head attention mechanism to capitalize on their combined strengths. Afterward, a detailed comparative result evaluation with other DL models is carried out to gauge robustness and accuracy in fault detection and localization. Comparative outcomes reveal that the Transformer models performed more remarkably in varying noise conditions, while recurrent models excelled in sequential data processing. The hybrid DL methods provided a balanced trade-off, improving fault detection and localization. The study utilizes several performance metrics such as accuracy, precision, recall, F1 score, MSE, RMSE, MAPE, and R2 errors to provide a comprehensive assessment. The fault detection without noise conditions attained an accuracy of 100%; however, there was a slight decrease (1.5%) in 20db noisy conditions. The proposed approach archives excellent performance in fault location tasks, such as MAPE:0.0308, MSE:1.6415, RMSE:1.2814, and R2:0.9982. The performance of the proposed approach is also validated with varying system operating conditions and comparative performance analysis with other state-of-the-art techniques. The findings offer valuable insights into the optimal model configurations and their practical applicability for reliable HVDC system protection, contributing to more effective and real-time fault diagnosis solutions.