The growing complexity of financial markets has made personal asset management a major challenge for individual investors. While robo-advisors and digital platforms aim to simplify investment decisions, they often use fixed allocation rules and miss important factors like behavioral risk, long-term stability, and market sentiment. This paper presents FinFlow, a Personal Asset Management System that combines machine learning, sentiment analysis, and financial optimization into one framework. The system is delivered through a web-based interface, offering investors personalized recommendations and an interactive dashboard. Evaluation using simulated user profiles and historical market data shows that FinFlow creates allocations that match investor preferences, reduces volatility through diversification, and increases responsiveness to sentiment-driven changes. The findings show the potential of AI-driven systems to fill the gap between the tools used by institutions and the needs of retail investors.

Atrial Fibrillation (AF) is among the leading arrhythmias with high occurrences of stroke, heart failure, and death. AF requires accurate and prompt identification to prevent serious outcomes of the disease, especially since AF can intermittently occur without symptoms in many patients. Electrocardiogram (ECG) is currently the principal method of identifying AF; however, its automated identification is complicated by noise, interpatient differences, and transient episodes of arrhythmia.In the past years, there has been a growing interest in the application of machine and deep learning for the detection of AF from the ECG signal due to the increased availability of larger ECG datasets. Although convolutional and recurrent neural networks have achieved superior performance, the majority of the existing solutions focus on the analysis of the monomodal ECG signal and have less interpretability and cross-platform performance. The article gives an extended review of the existing methods for AF detection, most of which were discussed, including signal processing techniques for the detection of AF, deep learning approaches for ECG signal analysis, and the new emerging methods for AF detection that take advantage of more medical information. In this review, the idea is to identify key research gaps and inspire the development of more robust, interpretable, and clinically relevant AF detection frameworks.

The unprecedented rise in the growth of Internet of Things has created an unprecedentedly enormous network of connected devices exchanging sensitive data. This huge transmission of security sensitive data poses several security issues which can be seldom addressed by conventional machine learning methodologies, opening an avenue for decentralized learning. Here, Federated learning emerges as a problem solver by letting several devices train models at different clients eradicating the requirement to move raw data from IoT devices to a centralized server, addressing security concern. This study examines the viabilities of federated learning within Internet of Things, suggesting a federated client-server architecture explicitly designed for capacity-constrained IoT devices. Current research work presents different aggregation techniques, such as FedAdam, FedAvg, and FedAdagrad and their efficiencies in IoT environment. Experiments show that Federated learning yields results at par with centralized models while achieving cost effectiveness and data privacy. The obtained results demonstrate that federated learning has the potential to escalate the efficacy of IoT applications encompassing healthcare, industrial automation, and intelligent transportation systems.

Colon cancer remains one of the leading causes of cancer-related mortality worldwide, with delayed detection significantly impacting survival rates. The time-consuming nature of traditional histopathological analysis contributes to these diagnostic delays. This study leverages advances in artificial intelligence to improve both the speed and accuracy of colon cancer detection through automated analysis of tissue sample images. We propose a weighted ensemble model comprising four convolutional neural networks namely RegNet X, RegNet Y, Swin B, and VGG11, that enhances prediction accuracy by combining outputs from base models. The prime focus of the research is to devise a recall prioritized ensemble model capable of detecting various types of colon cancers, while also examining the possibility of the model to generalize beyond the training domain. The ensemble was trained using a dataset of 10,000 histopathological images of colon tissues containing both benign and malignant samples. We introduce a novel preprocessing pipeline incorporating histogram equalization and contrast stretching for image enhancement with a motive to improve diagnostic accuracy. Notably, the ensemble model demonstrated superior recall metrics compared to traditional ML approaches—a critical measurement in healthcare applications. The ensemble model proposed has a test accuracy of 99% and a recall of almost 100%. Further, the proposed model is also tested for lung cancer to validate its generalizability. The experimental results obtained establish the validation of proposed model in terms of generalizability, demonstrating strong performance.

IntroductionHistopathological tissue reveals natural radial and bilateral symmetry in glandular structures, which becomes progressively disrupted during malignant transformation. Leveraging this observation, this work presents a VGG16-based deep learning model enriched with symmetry-aware interpretation for early detection of Colon Adenocarcinoma. The traditional approaches are not straightforward enough and acts as “black boxes” diminishing their clinical adoption and acceptance in real-world scenario. Current research work uses the most recent breakthroughs in deep learning on medical imaging and integrates Explainable AI strategies such as LIME, SHAP, and Grad-CAM into the model to interpret how cancer-induced symmetry distortions influence model decisions.MethodsThis work is experimented on a balanced dataset of 10,000 histopathological scans, including 5,000 Colon Adenocarcinoma tissue samples and 5,000 Benign Colon Tissue samples. This research aims to shed light on how benign tissues preserve consistent symmetric glandular patterns; while cancerous samples exhibit pronounced asymmetry, irregular boundaries, and disrupted structural repetition. Authors further aim to quantify these differences using lightweight 2D symmetry indices, demonstrating a clear separation between normal and malignant tissues.Results and DiscussionCurrent research presents a highly precise model for the diagnosis of colon cancer using a VGG16 CNN that achieves an encouraging test accuracy of 99.85%. The model exhibited very high precision, recall, and F1-scores for both classes, normal and cancer, as demonstrated by the classification report. Among various XAI techniques, Grad-CAM demonstrated speed and scalability making it an appropriate choice for its large-scale deployment in healthcare. SHAP, though computationally costly, offered theoretical robustness and great insight. LIME was handy in local interpretability, especially convenient in debugging individual predictions.

Electronic Health Records (EHR) are vital to modern healthcare, offering more effective means of electronically managing and accessing patient medical records. Using blockchain technology, this EHR makes use of Ethereum smart contracts for access and decentralized storage to provide security, transparency, and the ability to manage patient medical records in a tamper-proof way. The Interplanetary File System (IPFS), in conjunction with Pinata, provides immutable data storage for medical files. This EHR system combines smart contract-based access, decentralized storage of patient information, and a Web3 interface to support safe wardship of patient medical records while enhancing security and reducing administrative burden. It also includes a convolutional neural network (CNN) machine learning algorithm to harness the patient’s potentially harmful internal kidney conditions. The end product is an intelligent, data-driven, and secure EHR system that increases patient confidentiality of health information in settings with limited resources.

Component or system failures are inevitable in the industrial sector, leading to production downtime and significant financial losses. Consequently, each manufacturing sector requires an automated and efficient predictive maintenance system to monitor operations and predict potential failure points. Timely identification of these failure points facilitates prompt repairs, thereby substantially reducing maintenance costs. In this research, the authors aim to develop an effective model for the predictive maintenance of air compressors, which are critical components in various industries, including automotive, aerospace, chemical manufacturing, electronics, and general manufacturing. To achieve this, the authors employ a deep copy stacking classifier that utilizes an ensemble model composed of Logistic Regression, Decision Tree, Support Vector Machine, K-Nearest Neighbors, and Naive Bayes as base models to assess the bearing status of the components. The experimental evaluation of the proposed model demonstrates impressive performance metrics, achieving accuracy, precision, and recall rates of 99.3

This paper presents a novel approach for hyperparameter optimization for the MobileNetV2 architecture using a genetic algorithm. The proposed approach aims to automate the hyperparameter tuning leading to performance enhancement. This automated approach conserves the computational overheads involved in traditional hyperparameter tuning methods. This automated method for hyperparameter tuning is the result of significant advancement in the domain of deep learning models and eventually offers a scalable and efficient solution for developing high-performance mobile applications. This understanding will surely aid authors to devise an intriguing solution to address the involved challenges. Authors have provided 2-step solution where first part proposes a novel genetic algorithm based hyperparameter optimization followed by creation of a lightweight deep learning architecture, the second step of the solution. Further, the authors also aim to devise a mobile application that widens the scope of real-life application of the case study. Here, authors have undertaken the case study of poultry disease identification to evaluate the effectiveness and efficiency of proposed solution.

In order to optimize an objective function, it is always needed to invest resources in terms of cost, time, and many more. The more amount of resources is invested; objective function is likely to be improved till a certain threshold value. However, there needs to be trade-off between optimization of objective function and invested resources establishing a state of equilibrium. The major concerning aspect of such problems is the conflicting nature of resources and objective function, necessitating a specialized approach. Authors in this research work aims to attempt a novel framework that optimizes multi-objective function and provides a Pareto-front solution. Pareto-front solution consists of set of all feasible and optimal solutions which can be achieved in constrained scenario. In order to validate the effectiveness of proposed framework, authors have experimented the proposed framework on agricultural dataset where the objective is to maximize the crop yield while maintaining sustainability. Here, sustainability refers to minimizing the usage of fertilizers and pesticides. Authors here have used Non-dominated Sorting Genetic Algorithm-II for achieving multi-objective optimization. The experimental evaluation obtained on considered dataset validates the outperformance of Non-dominated Sorting Genetic Algorithm-II over traditional methods. The ability to achieve optimal solution in constrained scenarios further opens its avenues for real world applications.

In the rapidly changing world of electronic commerce, knowing and predicting what makes people buy things is key to making money. This research tests out how effective machine learning can be in analyzing customer data sets collected from e- commerce platforms. The researchers use a dataset gleaned from the UC Irvine Machine Learning Repository, which provides multiple predictors about user behavior during site visits as well as an outcome variable denoting whether they made purchases or not. Given that there are imbalances between different classes' representation rates within this sample size, our aim is developing classification models for purchase intentions while still considering such discrepancies [1]. In other words - through this project we want to find out what variables contribute most towards buying something online among all others available. The most appropriate algorithm turns out to be Random Forests due its stable performance over various runs where many decision trees get built simultaneously using different subsets of training data; then each tree votes independently so final predictions come from multiple paths thus reducing bias caused by using single model alone. When Hyperparameters are tuned, models tend generalize better by controlling complexity reduce overfitting improve predictive accuracy. This study concludes with suggestions for online firms, indicating how they can make use of machine learning findings in order to promote sales directed at specific customers and maximize profits. In general, such an investigation contributes towards furthering knowledge on data-based methods for comprehending revenue creation within electronic commerce settings.

During the past few years, the need for transparency and interpretability has been intensified owing to significant advancements in data-driven models, leading to the emergence of Explainable Artificial Intelligence (XAI). Several traditional XAI approaches are prevalent; however, these have limited competence in interpreting dynamic relations. The current research aims to address this limitation by proposing a novel Ensemble SHapley Additive exPlanations (SHAP) framework that focuses on temporal weighting, causal inference, hierarchical attribution, and interpretability optimization referred to as TCHSHAP. TCHSHAP prioritizes current information over historical information by temporal weighting through exponential decay. Further, causal inference separates correlation from causality to gain practical insights. Additionally, hierarchical attribution allows insights at granular (region level) and aggregated levels (feature-group impacts). These approaches are integrated to achieve a more interpretable and explainable model. To validate the efficacy of the proposed model, we carry out an experiment on the crop yield dataset collected from Kaggle. Ahead of experimental evaluation, data preprocessing is performed using one-hot encoding. Data normalization is done by min-max scaling, and outliers are removed through the Interquartile range. For the sake of experimental evaluation, the authors used the SHAP XAI model for Random Forest. When assessing the efficacy of the proposed TCHSHAP model, it is observed that while the average prediction for traditional SHAP is 161.137, it escalates to 161.506 after incorporating temporal weighting and causal inference, advocating the effectiveness of employing temporal and causal significance. Additionally, during hierarchical attribution, it is observed that agricultural features have the strongest dominance over the target variable. This dominance is followed by geographical and environmental factors in order. Thus, the obtained results authorize the efficacy of the proposed approach towards enhancing the global and local interpretability, strengthening the user's trust in model predictions. The current work offers ways to improve transparency and interpretability without affecting model performance. The suggested model also enables interpretable and efficient regression modelling in complex, data-driven applications, enabling its widespread application in real-world settings.

The proliferation of social networking sites has significantly improved longdistance communication and fostered a more interconnected global community. However, it also allows intruders to engage in surveillance and fraudulent activities. There has been a significant increase in such instances involving fake profiles, especially in recent years. To detect such profiles, this work presents a novel iterative greedy approach for detecting counterfeit profiles on social media networks. The work has considerable societal implications, as most antisocial behaviors on social networks are primarily carried out through fake accounts. The proposed methodology is implemented on three popular datasets. During the experimental setup, the base classifiers used are Support Vector Machine, Decision Tree, Logistic Regression, Gaussian Naïve Bayes, and k-Nearest Neighbor. The optimized ensemble model utilizes two decision trees and one k-Nearest Neighbor. The yielded ensemble model is also compared with base classifiers and traditional ensemble models, and it is observed that the proposed ensemble model outperforms base classifiers and traditional ensemble models in terms of F-score and accuracy.

The study ventures into the changing environment of blockchain, artificial intelligence (AI), and cloud computing as key components of financial technology (FinTech). This research examines how these technologies are transforming the financial system, with a focus on intelligent finance investment including blockchain’s disruptive potential. It begins with a brief description of cloud computing, explaining how it promotes extensibility, formability & cost-efficiency inside financial Institutions. The focus then shifts to AI applications, including the impact of robot advisers, trading using algorithms as well as statistical research on the reinvention of investing processes. The tale probes further into complex architecture of blockchain technology, giving light on its ability to increase the security and openness of the financial system. It reveals linkages that open the path for novel methods and significant developments, emphasizing the convergence of cloud, AI & blockchain. As we explore these FinTech territories, the paper acknowledges the inherent obstacles and dangers while also providing thoughts on legislative and ethical implications. It examines emerging trends, predicting the long-term influence of novel innovations on the financial environment.

This study focuses on solving several problems associated with Indian agriculture, which is a lifeline for 64

The emergence of diabetes caused due to increased level of blood glucose has widely affected the health of mankind across the globe. This is caused due to insufficient production of insulin that causes glucose to enter into cells and produce energy. Also scarcity of insulin may lead to continuous circulation of glucose preventing it to enter the cells leading to diabetes. The most prevalent forms of the disease are type 1, type 2, and gestational diabetes. As per the analysis of National Institutes of Health (NIH) as per 2015, around 9.4

Nowadays, music is a vital part of life and appeals to people of all ages. Even while music helps calm the body and mind, listening to songs with violence, profanity, or explicit lyrics can be harmful to mental health, especially in the young generation. It has been observed that teenagers who listen to explicit and violent music exhibit higher levels of hostility and less prosocial conduct establishing the impact of music on the mindset of mankind. Unfortunately, as the music can be accessed in a multitude of ways, it becomes difficult for parents to track what kind of music their children are listening to. Resultantly, it requires to devise an automated system that could identify the explicit music. In current research, authors have created an automated system using Natural Language Processing to identify explicit content in music using machine learning. This can be achieved through scanning the song lyrics for any kind of references to violence, foul language, or other offensive content which can further be used to flag potentially harmful content. This can serve as a helping hand for parents by providing them with better insight into what their kids are listening to. Working in this direction, the authors in this paper have proposed a majority voting-based binary classifier based on explicit content in the song lyrics. During experimental evaluation, the performance of the proposed methodology is matched with state-of-the-art methods to establish its efficacy. Thus, current research aims to foster a more secure artistic atmosphere for children and teenagers.