Comparative Study on Heart Disease Prediction Using Feature Selection Techniques on Classification Algorithms

Dissanayake, Kaushalya; Johar, Gapar Md

doi:10.1155/2021/5581806

Cited by 64 publications

(28 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…DT with forward, backward and bidirectional feature selections are the highest accuracy 98.2. In a significant comparison, the accuracy of our proposed model is found to be better than the recent studies [33] and [34], which were 88.52 and 85.8, respectively.…”

Section: Resultscontrasting

confidence: 52%

Analysis of Cardiac Anomalies by Selection and Extraction of Features Using Machine Learning Methods

Das¹,

Pradhan²,

S.³

et al. 2022

INDJCSE

View full text Add to dashboard Cite

One of the main factors that contribute to a person's death throughout the world is cardiac disease nowadays. Hospitals have vast amounts of clinical data stored in biomedical devices and other systems. Understanding the facts and finding the most significant features that can increase the estimate accuracy that may cause heart disease is crucial. In this paper, a model has been proposed by using relevant characteristics of selected various feature selection, feature extraction, and machine learning classifiers. Filter and wrapper methods are used for feature selection, whereas principal component analysis (PCA) and linear discriminant analysis (LDA) are used for feature extraction. Both approaches were applied to the Cleveland heart disease dataset to get the relevant attribute subset. Once the attribute subset is selected, different machine learning classifiers, namely Logistic Regression (LR), Naïve Bayes (NB), Random Forest (RF), Extreme Gradient Boost (XGB), K-Nearest Neighbour (KNN), Decision Tree (DT), and Support Vector Classifier (SVC) were used to classify the presence and absence of heart disease. Finally, the top three classifiers with the most significant attributes and different feature selection/extraction techniques are discussed. The performance of the proposed model was also calculated using various performance evaluation metrics like accuracy, precision, recall and f1 score.

show abstract

Section: Resultscontrasting

confidence: 52%

Analysis of Cardiac Anomalies by Selection and Extraction of Features Using Machine Learning Methods

Das¹,

Pradhan²,

S.³

et al. 2022

INDJCSE

View full text Add to dashboard Cite

show abstract

“…Heart failure is a clinical syndrome characterized by a reduced ability of the heart to pump blood to other parts of the body or fill with blood [1], [2]. Heart failure leads to fatigue, shortness of breath, and poor quality of life.…”

Section: Introductionmentioning

confidence: 99%

Prediction of patient survival from heart failure using a cox-based model

Assegie¹,

Karpagam²,

Subramanian³

et al. 2022

IJEECS

View full text Add to dashboard Cite

The existing heart failure risk prediction models are developed based on machine learning predictors. The objective of this study is to identify the key risk factors that affect the survival time of heart patients and to develop a heart failure survival prediction model using the identified risk factors. A cox proportional hazard regression method is applied to generate the proposed heart failure survival model. We used the dataset from the University of California Irvine (UCI) clinical heart failure data repository. To develop the model we have used multiple risk factors such as age, anemia, creatinine phosphokinase, diabetes history, ejection fraction, presence of high blood pressure, platelet count, serum creatinine, sex, and smoking history. Among the risk factors, high blood pressure is identified as one of the novel risk factors for heart failure. We have validated the performance of the model via statistical and empirical validation. The experimental result shows that the proposed model achieved good discrimination and calibration ability with a C-index (receiver operating characteristic (ROC) of being 0.74 and a log-likelihood ratio of 81.95 using 11 degrees of freedom on the validation dataset.

show abstract

“…However, they are limited in that they use a "one size fits all" approach [101]. Various methods of scoring features can be used, such as Pearson Correlation Coefficient, Mutual Information, or analysis of variance (ANOVA) [100,102]. For continuous input and categorical output, the set of input and output used in this work, ANOVA F-Test is a suitable filter method [102].…”

Section: Filter Methods -Anova F Testmentioning

confidence: 99%

“…The ANOVA F-Test selects for features which have the highest ratio of variance between groups, and variance within groups. It checks for variances between all features in the feature group, and selects those that have distributions with minimal overlap [102]. By selecting features with minimal or no overlap, we can attempt to provide features to the model that it can more easily distinguish between two groups of data [102].…”

Section: Filter Methods -Anova F Testmentioning

confidence: 99%

See 1 more Smart Citation

Characterization and Classification of Fibrillar Collagen Networks using Gray-Level Texture Analysis

Poole¹

View full text Add to dashboard Cite

Biological tissues are a combination of cellular and acellular components. The extracellular matrix (ECM), or the acellular component of tissues, is often overlooked.Tracking changes in the ECM grants insight into how aging, diseases and treatments might affect tissue structures and cellular responses. Using gray-level texture analysis, we analyzed the architecture of fibrillar collagen, a major component of the ECM.Several gray-level textural features were extracted from second harmonic generation (SHG) images and used to characterize collagen network arrangements. This work provides a basis for a classification model aimed to track changes in the cellular microenvironment. This information is vital if we wish to continue exploration of how diseases develop in different tissues, and develop novel and effective treatments for different types of chronic diseases.Foremost, and most certainly, I would like to extend my deepest and sincere gratitude to my supervisor Dr. Leila Mostaço-Guidolin. Her guidance, effort, and determination throughout the course of my studies, and in the writing of this thesis, have been invaluable. Especially during this period of history we find ourselves: the COVID-19 Pandemic, where studies felt like an uphill battle. She helped me ride out the storm, and all the while I gained valuable experience and knowledge that I'm sure will be helpful in the bright future. More so than that, she has been a source of inspiration, in her work ethic, dedication to her profession, and seemingly tireless ability to find time for her students in the midst of her own busy schedule.While I may not have met them, or know them, as much as I would have liked to due to the restrictions of the pandemic, I would like to thank my lab group, the BioImaging and Tissue Eng Lab. From the rare, and very much enjoyable, in-person meeting, to the mid-week Zoom meetings, it was always a pleasure to connect. I would especially like to thank Natasha Kunchur for her advice while I was writing this thesis. I would like to extend my thanks to Carleton University as a whole, an institution I have grown to know rather well in the seven years I have spent studying within iii its walls (two from the comfort of home), for its financial support during this degree.To my friends, for their constant support during my studies, which once appeared like they would never finish. At the times where overwhelming stress and anxiety became too much to bear, they would always be there to provide an ear to hear my complaints, a shoulder to cry on, or provide a much needed distraction.Finally, to my family, without whose support this thesis would never have been completed. Words cannot express how much gratitude I have for your support, guidance and understanding over the last few years. Especially my parents, Christopher Poole and Judith Poole, to whom I dedicate this thesis.

show abstract

Comparative Study on Heart Disease Prediction Using Feature Selection Techniques on Classification Algorithms

Cited by 64 publications

References 19 publications

Analysis of Cardiac Anomalies by Selection and Extraction of Features Using Machine Learning Methods

Analysis of Cardiac Anomalies by Selection and Extraction of Features Using Machine Learning Methods

Prediction of patient survival from heart failure using a cox-based model

Characterization and Classification of Fibrillar Collagen Networks using Gray-Level Texture Analysis

Contact Info

Product

Resources

About