Description:FIELD OF THE INVENTION
This invention relates to healthcare analytics and medical decision support systems. More specifically, it concerns a system and method for early detection of heart attacks using tuned machine learning and optimized artificial neural network (ANN) algorithms with advanced optimization techniques to improve accuracy, generalisation, and interpretability of predictions from multi-feature clinical data.
BACKGROUND OF THE INVENTION
Heart attacks remain one of the leading causes of death worldwide, often occurring suddenly and without sufficient early warning. Traditional diagnostic methods rely heavily on symptomatic evaluations, ECG interpretations, or manual risk scoring systems, which may fail to detect underlying cardiovascular issues at an early stage. Existing AI-based approaches, while promising, frequently suffer from limitations such as low generalization capability, suboptimal accuracy, lack of personalization, and poor interpretability—primarily due to static models, inadequate data integration, and manual parameter tuning. There is a pressing need for an intelligent, accurate, and proactive solution that can predict heart attack risk using diverse clinical data while addressing the shortcomings of current models. This research aims to solve this problem by developing a tuned machine learning system and an optimized Artificial Neural Network (ANN) framework using advanced optimization techniques like Particle Swarm Optimization (PSO) and Sharpness-Aware Minimization (SAM), enabling early detection, improved accuracy, and clinically interpretable results.
US2024366138A1: A method for early detection of a heart attack in a subject. The method includes acquiring a plurality of clinical symptoms from the subject, acquiring a gender of the subject, acquiring an age of the subject, acquiring a raw ECG signal from the subject, generating an averaged ECG signal from the raw ECG signal, acquiring a plurality of ECG features from the averaged ECG signal, designing a fuzzy inference system based on a set of rules associated with the plurality of clinical symptoms, the gender, the age, and the plurality of ECG features, and determining an occurrence of the heart attack utilizing the fuzzy inference system.
US4974598A: The Heart State Analyzer (HSA) is a system and method in medical non-invasive electrocardiographic (EKG) analysis of human heart beats for the early detection of certain heart diseases in which a large number of electrodes, for example, 32 to 64, are attached on the chest, back and sides of the patient, i.e., "body surface". The electrical signals detected by the electrodes ae converted to digital data, treated to remove muscle artifact and other noise, and then analyzed mathematically to determine the presence or absence of abnormal body surface potential distributions, or of unusual beat morphologies, compared statistically to the self-norm "typical beat" of the patient and also compared to a data base compiled from comparable normal population groups. The results of the statistical analysis are displayed as topographical maps of the body surface, color coded to represent the presence of significant derivations from the norms, defined as "abnormality", i.e., abnormal spatio-temporal patterns of voltages on the body surface, or as waveshape or histogram displays of features, similarly Z-transformed and color coded. Discriminant functions, stored in the heart state analyzer, estimate the relative probability of various cardiac pathologies.
Heart attacks remain a leading cause of death worldwide, often occurring without sufficient warning. Traditional diagnostic methods rely on symptomatic evaluations, ECG interpretations or manual risk scoring, which may fail to detect underlying cardiovascular issues at an early stage. Existing AI-based approaches often use static models with inadequate hyperparameter tuning, poor generalisation and limited interpretability. The present invention addresses these limitations by providing a tuned machine learning system and an optimized ANN framework employing Particle Swarm Optimization (PSO), Sharpness-Aware Minimisation (SAM) and belief-driven optimisation to deliver highly accurate, interpretable, and proactive heart attack risk prediction.
SUMMARY OF THE INVENTION
This summary is provided to introduce a selection of concepts, in a simplified format, that are further described in the detailed description of the invention.
This summary is neither intended to identify key or essential inventive concepts of the invention and nor is it intended for determining the scope of the invention.
The invention provides an intelligent and scalable system for early detection of heart attack risk using multi-feature clinical data. Patient data such as vitals, medical history and lifestyle factors are pre-processed and fed into an optimized ANN model whose hyperparameters are tuned automatically by a Particle Swarm Optimisation engine.
Sharpness-Aware Minimisation is incorporated into the training process to ensure the model generalises well to unseen patient data. Belief-driven optimisation enhances robustness in uncertain or borderline cases. An explainable AI module provides transparency by showing which factors most influenced each prediction.
The system outputs a risk score indicating the likelihood of a future heart attack. It can be deployed in hospitals, mobile health applications or wearable devices, enabling clinicians and patients to receive proactive alerts and make informed decisions. This combination of advanced optimisations within an ANN framework achieves a level of precision and generalisation not present in existing solutions.
To further clarify advantages and features of the present invention, a more particular description of the invention will be rendered by reference to specific embodiments thereof, which is illustrated in the appended drawings. It is appreciated that these drawings depict only typical embodiments of the invention and are therefore not to be considered limiting of its scope. The invention will be described and explained with additional specificity and detail with the accompanying drawings.
The proposed invention is the Early Detection of Heart Attacks using Tuned Machine Learning and Optimized ANN Algorithms—offers an intelligent and scalable solution to the problem of timely and accurate heart attack prediction. Unlike existing systems that rely on static models or rule-based logic, this methodology uses Artificial Neural Networks (ANNs) trained on comprehensive clinical datasets, including patient vitals, medical history, and lifestyle factors. To improve model performance and adaptability, advanced optimization techniques like Particle Swarm Optimization (PSO) are employed to fine-tune hyperparameters such as learning rate, number of neurons, and epochs. Additionally, Sharpness-Aware Minimization (SAM) is integrated into the training process to ensure the model generalizes well to unseen patient data by avoiding sharp minima in the loss landscape. The model is further enhanced with belief-driven optimization, making it robust in uncertain or borderline cases. The system processes patient data, runs predictions through the optimized ANN, and outputs a risk score indicating the likelihood of a future heart attack. It also supports interpretability through tools like SHAP or LIME, which help clinicians understand which factors most influenced the prediction. This approach not only improves accuracy, recall, and generalizability but also enables proactive clinical decision-making. As a result, the system can be deployed in hospitals, mobile health apps, or wearable devices, providing a powerful tool for early intervention and potentially saving lives.
BRIEF DESCRIPTION OF THE DRAWINGS
The illustrated embodiments of the subject matter will be understood by reference to the drawings, wherein like parts are designated by like numerals throughout. The following description is intended only by way of example, and simply illustrates certain selected embodiments of devices, systems, and methods that are consistent with the subject matter as claimed herein, wherein:
FIGURE 1: SYSTEM ARCHITECTURE
The figures depict embodiments of the present subject matter for the purposes of illustration only. A person skilled in the art will easily recognize from the following description that alternative embodiments of the structures and methods illustrated herein may be employed without departing from the principles of the disclosure described herein.
DETAILED DESCRIPTION OF THE INVENTION
The detailed description of various exemplary embodiments of the disclosure is described herein with reference to the accompanying drawings. It should be noted that the embodiments are described herein in such details as to clearly communicate the disclosure. However, the amount of details provided herein is not intended to limit the anticipated variations of embodiments; on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the scope of the present disclosure as defined by the appended claims.
It is also to be understood that various arrangements may be devised that, although not explicitly described or shown herein, embody the principles of the present disclosure. Moreover, all statements herein reciting principles, aspects, and embodiments of the present disclosure, as well as specific examples, are intended to encompass equivalents thereof.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments. As used herein, the singular forms “a",” “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises,” “comprising,” “includes” and/or “including,” when used herein, specify the presence of stated features, integers, steps, operations, elements and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and/or groups thereof.
It should also be noted that in some alternative implementations, the functions/acts noted may occur out of the order noted in the figures. For example, two figures shown in succession may, in fact, be executed concurrently or may sometimes be executed in the reverse order, depending upon the functionality/acts involved.
In addition, the descriptions of "first", "second", “third”, and the like in the present invention are used for the purpose of description only, and are not to be construed as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Thus, features defining "first" and "second" may include at least one of the features, either explicitly or implicitly.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which example embodiments belong. It will be further understood that terms, e.g., those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
The proposed invention is the Early Detection of Heart Attacks using Tuned Machine Learning and Optimized ANN Algorithms—offers an intelligent and scalable solution to the problem of timely and accurate heart attack prediction. Unlike existing systems that rely on static models or rule-based logic, this methodology uses Artificial Neural Networks (ANNs) trained on comprehensive clinical datasets, including patient vitals, medical history, and lifestyle factors. To improve model performance and adaptability, advanced optimization techniques like Particle Swarm Optimization (PSO) are employed to fine-tune hyperparameters such as learning rate, number of neurons, and epochs. Additionally, Sharpness-Aware Minimization (SAM) is integrated into the training process to ensure the model generalizes well to unseen patient data by avoiding sharp minima in the loss landscape. The model is further enhanced with belief-driven optimization, making it robust in uncertain or borderline cases. The system processes patient data, runs predictions through the optimized ANN, and outputs a risk score indicating the likelihood of a future heart attack. It also supports interpretability through tools like SHAP or LIME, which help clinicians understand which factors most influenced the prediction. This approach not only improves accuracy, recall, and generalizability but also enables proactive clinical decision-making. As a result, the system can be deployed in hospitals, mobile health apps, or wearable devices, providing a powerful tool for early intervention and potentially saving lives.
This proposed system uniquely combines Sharpness-Aware Minimization (SAM), belief-driven optimization, and Particle Swarm Optimization (PSO) within an Artificial Neural Network framework to enable highly accurate, interpretable, and early prediction of heart attacks from multi-feature clinical data—offering a level of precision and generalization not present in existing solutions.
The invention comprises an input interface for receiving multi-feature patient data, including demographics, vital signs, medical history, diagnostic test results, and lifestyle factors.
A data preprocessing module cleans, normalises, and structures the incoming data to ensure quality and consistency for model input. This includes handling missing values, scaling continuous variables, and encoding categorical data.
A feature integration engine combines clinical and personal health data into a unified representation, enabling the model to capture complex nonlinear relationships between factors.
An artificial neural network (ANN) serves as the core predictive model. It is configured with multiple hidden layers, activation functions and output nodes representing heart attack risk.
A Particle Swarm Optimisation (PSO) module automatically tunes key hyperparameters of the ANN such as learning rate, number of neurons and number of epochs, eliminating manual trial-and-error and improving convergence.
Sharpness-Aware Minimisation (SAM) is integrated into the training process to prevent overfitting and enhance generalisation by avoiding sharp minima in the loss landscape.
Belief-driven optimisation further refines the model’s predictions in uncertain or borderline cases, increasing robustness and reducing misclassification.
The system includes an explainable AI (XAI) module, such as SHAP or LIME, which interprets the ANN’s predictions by identifying the contribution of each feature to the risk score, thereby increasing clinician trust and facilitating personalised interventions.
A risk scoring module computes a heart attack risk probability for each patient based on the ANN’s output and presents it in an easily interpretable format.
An alert generation module issues notifications or recommendations when risk thresholds are exceeded, enabling proactive clinical intervention.
The system can be deployed as a cloud service, on-premise hospital server, or integrated with mobile and wearable devices for continuous monitoring.
Security and privacy safeguards ensure that patient data is encrypted, anonymised where necessary, and handled in compliance with healthcare data standards.
Periodic retraining of the model with new data ensures that performance remains high across evolving patient populations and changing medical practice patterns.
This architecture provides a comprehensive, proactive, and interpretable heart attack prediction tool, reducing false positives and missed detections compared to existing solutions.
The invention can be adapted to other cardiovascular conditions or extended to general health risk prediction by retraining on appropriate datasets.
BEST METHOD OF WORKING
The preferred embodiment deploys the system as a cloud-based platform integrated with hospital electronic health records. Patient data is securely uploaded and preprocessed. The optimized ANN model, trained using PSO, SAM and belief-driven optimisation, predicts heart attack risk scores. The explainable AI module displays factor contributions, while an alert system notifies clinicians of high-risk patients. This configuration achieves high predictive accuracy, recall, and F1-score with low latency, enabling proactive and personalised heart attack prevention.
, Claims:1. A system for early detection of heart attack risk comprising:
an input module configured to receive multi-feature patient data including demographics, vital signs, medical history, diagnostic test results and lifestyle factors;
a preprocessing module configured to clean, normalise and structure the data;
a feature integration engine configured to combine clinical and personal health data into a unified representation;
an artificial neural network (ANN) module configured to predict heart attack risk;
a Particle Swarm Optimisation module configured to automatically tune hyperparameters of the ANN;
a Sharpness-Aware Minimisation module configured to enhance model generalisation;
a belief-driven optimisation module configured to improve robustness in uncertain cases;
an explainable AI module configured to interpret and display feature contributions to predictions; and
an output module configured to present risk scores and alerts to a user.
2. The system as claimed in claim 1, wherein the preprocessing module handles missing values, scales continuous variables and encodes categorical data for model input.
3. The system as claimed in claim 1, wherein the Particle Swarm Optimisation module tunes learning rate, number of neurons and number of epochs of the ANN to improve convergence.
4. The system as claimed in claim 1, wherein the Sharpness-Aware Minimisation module prevents overfitting by avoiding sharp minima in the loss landscape.
5. The system as claimed in claim 1, wherein the explainable AI module uses feature attribution techniques to provide transparency of predictions.
6. A method for early detection of heart attack risk comprising:
receiving multi-feature patient data including demographics, vital signs, medical history, diagnostic test results and lifestyle factors;
preprocessing the data to clean, normalise and structure it;
combining the data into a unified representation;
predicting heart attack risk using an artificial neural network;
automatically tuning hyperparameters of the neural network using Particle Swarm Optimisation;
enhancing model generalisation using Sharpness-Aware Minimisation;
improving prediction robustness using belief-driven optimisation;
interpreting predictions using an explainable AI module; and
outputting risk scores and alerts to a user.
7. The method as claimed in claim 6, wherein preprocessing includes handling missing values, scaling continuous variables and encoding categorical data.
8. The method as claimed in claim 6, wherein hyperparameters including learning rate, number of neurons and number of epochs are tuned automatically by Particle Swarm Optimisation.
9. The method as claimed in claim 6, wherein explainable AI interprets and displays contributions of individual features to each risk prediction.
10. The method as claimed in claim 6, wherein risk scores and alerts are integrated into hospital electronic health records, mobile health applications or wearable devices for continuous monitoring.