Project Phase-01.pptx guj.fjfjyfjfhjjfgj

fhfjfjfjhf

1. PES COLLEGE OF ENGINEERING MANDYA DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING PROJECT PHASE- I “ MACHINE LEARNING APPROACH FOR THE PREDICTION OF CARDIAC ARR HYTHMIA ” UNDER THE GUIDANCE OF: Mrs. Deepika . B.E., MTech. Assistant Professor Department of CS&E PRESENTED BY: Nithyashree M P (4 PS19CS066 ) Prakruthi H S (4 PS19CS073 ) Ra kshitha H N (4 PS19CS081 ) S hreekala M K (4 PS19CS102 ) 7th Sem, Department of CS&E

2. Introduction:- Cardiac Arrhythmia is a condition in which the heartbeat is irregular, either it is too fast or too slow. Many types of Arrhythmia have no symptoms. When symptoms are present these may include palpitations or feeling a pause between heartbeats. In more serious cases there may be lightheadedness, passing out, shortness of breath, or chest pain. While most types of arrhythmia are not serious, some predispose a person to complications such as stroke or heart failure. Others may result in cardiac arrest. Arrhythmia affects millions of people. About half of the deaths are due to cardiovascular diseases. About 80% of sudden cardiac death is the result of ventricular Arrhythmias. Arrhythmias may occur at any age but are more common among older people.

3. Risk Factors of Cardiac Arrest:- Number of Factors can cause the heart to work incorrectly.They include, Alcohol Abuse Diabetes Drug Abuse Excessive Coffee Consumption Heart disease like congestive heart failure Hypertension(High Blood Pressure) Hyperthyroidism(An overactive thyroid gland),etc…..

4. LITERATURE SURVEY:- Sl.No Title Author/s Algorithms used Contributions Limitations 01 Heart Disease Prediction using Machine Learning Techniques VV Ramalingam Ayantan Dandapath M.Karthik Raja [SRM Institute of Science &Technology] Naive Bayes Algorithm Support Vector Machine(SVM) K-nearest Neighbour(KNN) Decision Tree Algorithm Random Forest(RA) Ensemble Model Algorithms used here has huge scope in predicting cardiovascular diseases. Random Forest Algorithm is used to overcome the overfitting problem. For some data the algorithms work poorly due to overfitting. There is a need to research on how to handle high dimensional data and overfitting. Lot of research is needed to be done on the correct ensemble of algorithms to use for particular type of data 02 Comparison of Machine Learning Techniques for Hospitalization in heart failure patients Giulia Lorenzoni Stefano Santo Corrado Lanera Clara Minto Honaria Ocagli Poola De Paolis Franco Pisano Sabina Iliceto Decision Tree Algorithm Naive Bayes Random forest Logistic Regression Support Vector Machine(SVM) Helps in both clinical and economic stand points Number of techniques improve the heart failure diagnosis process such as extreme learning machine,heart disease classification etc. Improved performance based on medical history Cost of treatment is high. No improvement using machine learning techniques compare to traditional techniques Burden to hospital admission or readmission

5. Continued… Sl No Title Author Algorithms used Contributions Limitations 03 Machine Learning approach for sudden cardiac arrest prediction based on optimal heart rate variability features L.Murukesan Loganathan Ye Htut Murugappan M. Filter or wrapper methods Tree Bagger classifier Sequential Feature Selection Algorithm(SFS) Support Vector Machine(SVM) Probabilistic Neural Network(PNN) Algorithm here used for SCA prediction appears to be more sensible way of predicting SCA compared to others Ectopic beats detected Given a relatively humble dataset of 38 observations it is not sensible to generalize the findings for much larger populations 04 Heart Disease prediction using supervised machine Learning Algorithms. M Mamun Ali Bikash Kumar Paul Kawsar Ahmed Francis M Bui Julian M.W.Quinn Mohmmad Ali Moni K-Nearest Neighbour Decision Tree Algorithm Random Forest The algorithms performed extremely with more accuracy Data mining techniques can do the job efficiently at a very low cost using a classification algorithm which plays a key role in clinical research ML and data mining based approaches to prediction to prediction and disease would be of great clinical utility but are highly challenging to develop.

6. Continued…. Sl No Title Authors Algorithms used Contribution Limitations 05 A hybrid deep model for automatic Arrhythmia classification based on LSTM recurrent networks Adeleh Bitarafan Afra Amini Mandich Soleymani Baghshah Hounidreza Khodajou Chokami Decision Tree Random Forest Ensemble Algorithm LSTM recurrent networks remits modelling the relation among different heartbeat in a sequence. Due to not considering any export knowledge and automatically segmenting and class We could carry out the experiment only under the weighted cross entropy loss 06 Deep multiscale fusion neural network for multi class Arrhythmia detection Ruxin wang Jianping Fan and Li Decision Tree Random Forest Ensemble Algorithm Deep multi scale fusion(DMSF) CNN architecture is used for capturing discriminative signal features from multi scales The experimental results doesnot convince the effectiveness of deep multiscale fusion network to other physiological signal analysis and process requirements other than multiscale,ECG signal features.

7. Continued … Sl No. Title Authors Algorithms used Contributions Limitations 07 Automated Arrhythmia detection from electrocardiogram signal using stacked restricted Boltzmann machine model Saroj Kumar Pandey Ram Jhangal Adithya VikramDev Pankaj Kumar Mishra 1D convolutional neural networks(CNN) Myocardial infarction(MI)classification Restricted Boltxmann Machine(RBM) Database: MIT-BIH Arhythmia Database(AD) The proposed methodologies performance is evaluated for every category of ECG signal The classifer has achieved an overall accuracy of 99.61%. It requires specialized hardware 08 A framework for cardiac Arrhythmia detection from IOT based ECG’s Jiayuan He Jia Rong Le Sun Hua Wang Yanchuan Zhang & Jiangang Ma Decision Tree Random Forest Ensemble Algorithm Dynamic heart beat classification with adjusted features(DHCAF) and multi channel heartbeat convolution neural network(CNN) Heartbeat rhythms are not integrated well to the network and also it can be easily by the other learned features.

8. Existing Systems:- Existing monitoring systems for ECGs record utilize algorithms to determine changes in cardiac rhythm. However, accurate identification of arrhythmias is known to be challenging even for medical professionals, and requires considerable medical expertise. Limitations of Existing System:- Accuracy is Low. In some existing systems heartbeat rhythms were not integrated well into the network. Some of the systems do not detect arrhythmia.

9. Current Approach:- To develop a system that detects and classifies cardiac arrhythmia This system aims at using different machine learning algorithms like Naive Bayes, SVM, and Random Forests for predicting and classifying arrhythmia into different categories. Using the ensemble method to identify the cardiac arrhythmia classification.

10. Flow of Project:- DATASET PRE- PROCESSING FEATURE EXTRACTION CLASSIFICATION EVALUATION

11. Software Requirements :- Operating system : macOS, Windows XP/7 or higher version. Coding language : python (>=python 3.3). IDE : Jupyter Notebook.

12. Hardware Requirements:- Processor: State-of-the-art i3 Ram: 4GB Hard Disk: 500GB Input device: Standard Keyboard and Mouse. Compact Disk: 650Mb. Output device: High-Resolution Monitor .

13. THANK YOU