INTERNATIONAL JOURNAL OF NOVEL RESEARCH AND DEVELOPMENT International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2456-4184 | Impact factor: 8.76 | ESTD Year: 2016
Scholarly open access journals, Peer-reviewed, and Refereed Journals, Impact factor 8.76 (Calculate by google scholar and Semantic Scholar | AI-Powered Research Tool) , Multidisciplinary, Monthly, Indexing in all major database & Metadata, Citation Generator, Digital Object Identifier(DOI)
A competent neural network with convolutional layers is reported in this article, ideal for recognizing extremely
unexpected clinical data. The National Health and Nutritional Examination Survey (NHANES) is the source of the carefully
collected data used to estimate rates of Coronary Heart Disease (CHD). When applied to this kind of data, typical machine learning
algorithms usually have permanent class mismatch issues, even after class-specific weights have been modified. However, our
updated two-layer CNN performs as well in each group, suggesting that it is resilient to such modifications. In the scenario of a
very unequal dataset, we apply a two-pronged technique to acquire greater accuracy for both positive (actual CHD predictions)
and negative (healthy cases) classes on an increasing test sample size. First, we employ the Least Absolute Shrinkage and Selection
Operator (LASSO) to assess feature weights. This is replaced by a majority-voting mechanism for identifying significant qualities.
A fully connected layer is a vital step prior to these essential qualities traveling across succeeding convolutional layers for
comparison. To further increase overall classification accuracy, we provide an epoch-specific training strategy that replicates a
simulated annealing process. Our recommended CNN architecture obtains an extraordinary classification accuracy of 77% for
positive CHD cases and 81.8% for negative instances on the testing data, which represents for 85.70% of the overall dataset,
despite the frequent class conflict in the NHANES dataset. This result implies the likely generalizability of the recommended
strategy to other healthcare research with comparable feature ordering and discrepancies. Although our CNN model's memory
scores are comparable to those of other machine learning approaches like SVM and random forest, our model is better at
identifying negative (non-CHD) events. The recommended strategy may enhance medical treatments, minimize diagnostic
charges, and develop study tools by putting a smart diagnostic system inside the healthcare framework. Notably, our model's total
accuracy (79.5%) outperforms the accuracy of the SVM and random forest models separately. Data like as sensitivity, test
accuracy, recall, and Area Under the Curve (AUC) are created using the CNN algorithm.
Keywords:
Cite Article:
"Predictive Capability of a Condensed CNN Model in Discovering Patterns of Coronary Heart Disease", International Journal of Novel Research and Development (www.ijnrd.org), ISSN:2456-4184, Vol.9, Issue 2, page no.d144-d157, February-2024, Available :http://www.ijnrd.org/papers/IJNRD2402316.pdf
Downloads:
000118770
ISSN:
2456-4184 | IMPACT FACTOR: 8.76 Calculated By Google Scholar| ESTD YEAR: 2016
An International Scholarly Open Access Journal, Peer-Reviewed, Refereed Journal Impact Factor 8.76 Calculate by Google Scholar and Semantic Scholar | AI-Powered Research Tool, Multidisciplinary, Monthly, Multilanguage Journal Indexing in All Major Database & Metadata, Citation Generator
Facebook Twitter Instagram LinkedIn