Architectural Engineering Students Explore Mixed-Use Design Concepts at Qaiwan Towers
Dr. Wafaa Mustafa Hameed is a Lecturer from the Computer Science Department at Cihan University Sulaimaniya who has Published a high-impact article in Expert Systems with Applications journal with (Impact Factor 7.5, Q1 Scopus-indexed, Clarivate and Cite Score 13.80). The paper, titled “Enhancing accuracy of diabetes diagnosis system using instance-wise feature importance aware imputer and tabnet deep neural classifier”, This paper presents a diabetes diagnosis model by presenting a novel prediction and instance-wise feature importance aware imputation technique, and explainable Tabnet classifiers. that surpasses the existing imputation techniques by a large margin and achieves competitive results with the ideal but impractical MCA imputer.
The method presented in this research called PFAI-Tabnet enriched with outlier identification and treatment technique. Also, the research handled imbalanced nature of the task by examining two resampling techniques: SMOTETomek and SMOTEENN. The results show that SMOTEENN is effective for producing balanced results in a Diabetes recognition system. The application of an outlier handling strategy greatly improved the performance of PFAI-Tabnet on the Pima Diabetes dataset from 89.61 % to 93.51 %, or about a 4.9 % increase. The improvement shows how effective the management of outliers is in refining model performance, where the training data becomes more representative and less influenced by abnormal values.
Comparisons with state-of-the-art methods on both evaluated datasets confirm that PFAI-Tabnet outperforms these methods by a large margin on the variety of classification metrics
Read the full article here:
https://www.sciencedirect.com/journal/expert-systems-with-applications/vol/270/suppl/C
Related posts
