Bhartiya Skill Development University

The systematic process of locating and evaluating structural damage to structures brought by seismic occurrences is known as earthquake damage detection, and it serves to both ensure public safety and direct efforts toward rebuilding or rehabilitation. The optimal use of remote sensing for earthquake building damage identification is hampered by misclassification brought by particulate interference and shadowing impacts. To overcome this problem, this study investigates the use of data science tools in the estimation of earthquake damage, to improve the precision and effectiveness of post-disaster evaluations. The deep learning (DL) approach was used in the proposed method, Advanced Barnacles Mating Optimization-Driven Deep Belief Networks (ABMO-DBN),for earthquake building damage detection. A comprehensive earthquake dataset was gathered. This study is experimented on the Python platform. The suggested approach is contrasted with the various approaches that are currently in use, the findings show the suggested technique achieves enhanced performance in average precision (93.99%), precision (95%), F1-score (92.55%), and recall (90%). The objective of the research is to forecast the level of structural damage to several seismic characteristics, such as depth, distance from the epicenter, and magnitude. Predictive models that can direct resource allocation and emergency response activities can be developed through the integration of historical records and geographic data. The results demonstrate how data-driven methods may enhance disaster management plans by providing a quicker and more accurate assessment of earthquake effects, which is essential for minimizing losses and guaranteeing a speedy recovery.

Residential power grids are required to integrate energy storage devices and renewable energy sources of growing need for renewable energy. However, managing the variable nature of these energy sources while ensuring system reliability and cost efficiency remains a significant challenge. Current optimization techniques often struggle to address the dynamic fluctuations of renewable energy generation and the interplay between diverse energy sources, leading to inefficiencies and increased operational costs. Hence the multi-source energy management system is enhanced by implementing a novel intelligent Kepler optimization (IKO) framework in residential systems. Mathematical model for energy resources such as solar energy, wind energy, battery storage and grid power were developed for thermal and electrical appliances such as washing machines, television (TV), and microwave oven. The objective of the research is to implement an exact solution for the energy consumption problem that minimizes electricity costs in a residential power system, and optimizing the operation modes of various energy sources while coordinating their interaction for efficient energy management. The results were produced based on computing time, power consumption, and energy efficiency for washing machines (95sec)(480W)(92%), TV(75sec)(180W)(88%), and microwave oven (80sec) (1100W) (82%). The proposed method provides a robust and sustainable solution for energy management in residential power systems, effectively optimizing the multiple energy source and improving overall system efficiency.

Building constructions that are based on sophisticated mathematical models for seismic activity assessment use complex simulations and algorithms to determine how buildings react to earthquakes. By examining possible effects and weaknesses, this method assists engineers in designing safer, more robust structures by forecasting structural performance under seismic challenges. Seismic activity evaluation in building structures is limited to its reliance on simulation models and historical earthquake data, which cannot accurately represent the full spectrum of seismic occurrences, potentially leading to underestimating the actual danger. The Elephant Herding Optimized Dynamic Recurrent Neural Network (EHO-DRNN) combines data with adaptive algorithms to overcome constraints in seismic activity evaluation. By continually updating and improving models, EHO-DRNN improves predicted accuracy and better captures a variety of seismic events. The earthquake data are gathered, and the classified technique is using suggested approach EHO-DRNN. As a result, the earthquake data had to be generated. The proposed method is also compared to the other traditional algorithms, and this research experimented with the Python platform. The findings demonstrate that the suggested strategy outperformed the lower error in R (0.99994), R2 (0.99983), MAE (0.01418), and RMSE (0.05209). High-performance mathematical models enhance seismic activity evaluation in building structures, enhancing structural resilience, safety precautions, and decision-making in earthquake design.

Stock market prediction based on textual information, has become increasingly important due to the vast amount of financial news and social media content influencing market behavior. This research aims to enhance stock market trend prediction by leveraging sentiment analysis of social media data using an Intelligent Grid Search Adaptive k-Nearest Neighbors (IGS-AKNN). The dataset consists of 30,000 tweets categorized as positive, negative, and neutral. Data preprocessing is performed utilizing Natural language processing (NLP) approaches and feature extraction is done using the term frequency-inverse document frequency (TF-IDF) method. The findings indicate that the proposed model outperforms traditional methods in terms of recall (0.852), accuracy (0.870), F1-score (0.875) and precision (0.866) with lower error metrics, including MAPE (0.015342) and MSE (0.005628). This delves into the efficiency of the IGS-AKNN method in predicting the destiny of share price fluctuation based on social media sentiment.

In corporate investment, predicting a company's financial failure is a crucial task that has significant implications for regulators, investors, and other stakeholders. Because of the complexity and huge amount of financial data, precise forecasting requires sophisticated techniques. This research aims to enhance the early detection of financial failure by designing a sophisticated predictive model named Brown Bear Search-driven Intelligent Support Vector Machine (BBS-ISVM) to integrate and analyze company-specific financial indicators, macroeconomic factors, and operational metrics. The dataset consists of historical financial data from diverse industries, including balance sheets, covering metrics, revenue, and net income. Data preprocessing involves z-score normalization, and Principal Component Analysis (PCA) is utilized for feature extraction. The findings show that the suggested model performs noticeably better than traditional methods, attaining higher performance measures, such as F1-score (86.5%), recall (86%), accuracy (80%), and precision (85.5%). Additionally, the approach offers useful insights into the major financial metrics affecting a business's possibility of failure. These results demonstrate a hybrid machine learning model's potential as a reliable tool for identifying financial failures early and providing helpful assistance for risk control and investment strategy decision-making.

The world is rapidly transitioning to a cashless culture today, which has led to a huge rise in credit card usage. However, the flip side of this is that there is an increase in fraudulent activity for this reason, both the card-issuing institutions and cardholders need to create a methodical fraud detection system. This study presents a novel method for CCFD using a Fruit fly driven weighted random forest (FF-WRF). This paper uses a Kaggle credit card transaction dataset and employs data cleaning, linear Discriminant Analysis for feature extraction and FFA to improve model accuracy. Furthermore, fraud detection accuracy is increased with an upgraded Random Forest method that incorporates feature weighting and subspace sampling. Based on a comparative analysis, FF-WRF outperforms conventional techniques such as RF, SVM, and LR, with rates of accuracy (98.3%), precision (87%), recall (95%), and F1-score (89%) respectively. These outcomes demonstrate the superior performance of FF-WRF in credit card fraud detection, providing a reliable way to reduce false positives and negatives.

Hidden feelings are unveiled with facial micro expressions (FME) finding applications in security, psychoanalysis, and human-computer interaction. Thus, the aim of this work is to design a new ML algorithm that accurately detects FME in real-time, which finding is complex, accurate and fast enough to benefit fields such as psychology, security and human-computer interaction. The proposed approach employs Sooty Tern Second optimizer (ST-ERF) based Enriched Random Forest (ERF) for timely recognition of FME. The dataset consists of image data on FME, including various emotions and action units, with high temporal resolution, enabling accurate recognition of subtle, dynamic facial expressions. Bilateral filtering smoothes face regions without smoothing edges, and the high-pass filter helps highlight fine expression details. It extracts essential features by applying histogram of oriented gradient (HOG). ST helps in enhancing convergence by quick navigation of loss landscape and ERF detects spatial and temporal dynamics in sequential facial expression data. The suggested ST-ERF technique is highly accurate in real-time ME recognition, exceeding previous models in efficiency and precision.

Background: Visual arts education has been suggested to influence cognitive and behavioral outcomes, yet its specific effects on critical thinking and behavior regulation remain underexplored. The effect of visual arts education on these domains forms the topic of the present study, which employs a mixed approach research design. Aim: To assess the impact of visual arts education on the improvement of critical thinking and self-regulation in students. Methodology: A quasi-experimental approach was employed with 80 individuals, with 40 in the experimental group Report Phrase receiving education in visual arts and 40 in the control group receiving non-visual arts education. CTST and BRI assessment tools were used to acquire quantitative data about the participants, utilizing the independent and paired sample T-tests. Data collected from participants include interviews, focus group discussions, self-administered, and thematic analysis. A repeated measures ANOVA was used to analyse the data. Result: Statistical findings depicted improvements in the quantitative aspect; critical thinking skills improved from 70.0 to 75.3 with p<0.001 and behavior regulation improved from 65.0 to 69.2 with p<0.001 in the experimental-group in contrast to the control-group. The thematic analysis found striking increases in attention, impulse control, task completion, creativity, collaboration, affect expression, and self-esteem of experimental-group participants. Based on the findings of the Repeated Measures ANOVA, significant enhancements were observed in both domains for critical thinking and behavior regulation with effect size estimates of 0.40 and 0.30, respectively. Conclusion: The development of critical thinking and behavior control in students is greatly impacted by art instruction. Quantitative and qualitative data both support the utility of using visual arts interventions for cognitive and behavioral development and inform practice in art education.

In order to better understand the market and make well-informed decisions, data-driven marketing, or DDM, is a marketing strategy that uses artificial intelligence (AI) technology to gather and analyze vast volumes of data. This makes it possible for the business to provide clients with a tailored customer experience and react to their unique demands more efficiently. In order to establish a privileged, long-lasting connection between the consumer and the brand, MDD is also engaged in the implementation of data-driven loyalty plans. Customers respond to these tactics by exhibiting traits that help maintain this connection, such as considerate loyalty, a willingness to advocate for the brand, and a willingness to work with it in the long run. By doing this, the brand satisfies the increasing demands of the competition and accomplishes its goals of development, sustainability, and profitability.