Missing Data Imputation Method for Autism Prediction
Kamatchi Priya L1, Baranidharan C2
1Kamatchi Priya L, Associate Professor, New Horizon College of Engineering, Bangalore, India.
2Baranidharan C, AVEVA Solution.
Manuscript received on January 02, 2020. | Revised Manuscript received on January 15, 2020. | Manuscript published on January 30, 2020. | PP: 940-944 | Volume-8 Issue-5, January 2020. | Retrieval Number: D4551118419/2020©BEIESP | DOI: 10.35940/ijrte.D4551.018520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Missing data imputation is essential task because removing all records with missing values will discard useful information from other attributes. This paper estimates the performance of prediction for autism dataset with imputed missing values. Statistical imputation methods like mean, imputation with zero or constant and machine learning imputation methods like K-nearest neigh bour chained Equation methods were compared with the proposed deep learning imputation method. The predictions of patients with autistic spectrum disorder were measured using support vector machine for imputed dataset. Among the imputation methods, Deep learning algorithm outperformed statistical and machine learning imputation methods. The same is validated using significant difference in p values revealed using Friedman’s test.
Keywords: PTFE, Powder Metallurgy, Reinforcement, Toughness.
Scope of the Article: Service Discovery and Composition.