29 Tablet form and color design method based on support vector regression and swarm intelligence. (19th December 2016)

• Record Type: Journal Article
• Authors: Ding, Man; Chen, Hanning
• Is Part Of: Journal of investigative medicine. Volume 64(2016)Supplement 8
• Page Start: A10
• Page End: A11
• Publication Date: 2016-12-19
• Subjects: Clinical medicine -- Periodicals; Medicine -- Research -- Periodicals; Medicine; Research -- United States; Clinical medicine; Medicine -- Research; Periodicals; 616.075
• DOI: 10.1136/jim-2016-000328.29
• Languages: English
• ISSNs: 1081-5589

Abstract

Abstract : Objectives: The emotional effect of a tablet's form and color on patients can affect their choice of drug, and their confidence in the effect of treatment. In order to improve the drug's treatment effect, it is necessary to design the tablet's form and color to meet the patients' emotional needs. This paper proposes a method for designing the form and color of tablets based on support vector regression and a swarm intelligence algorithm. Methods: First, the method decomposes the design elements of the tablet, and assigns category labels and fuzzy affiliation values to tablet samples in order to build the classification model. The patients' perceptions of tablet samples can be obtained through their subjective evaluation. The mapping model between design elements and patients' perceptual images can then be built using a support vector regression model. Finally, the design scheme is optimized using a swarm intelligence algorithm. Results: The proposed method can assist designers to quickly identify a satisfying design scheme despite limited options, and can increase the success rate of tablet design and improve treatment effect. Conclusions: The proposed method can effectively produce a satisfying design scheme while accounting for different image preferences caused by individual differences. Therefore, this method can provide a valid approach for large-scale design and theoretical support for tablet design. Acknowledgments: This research was financially supported by the National Natural Science Foundation of China (Grant No. 51575158) and Hebei Province Natural Science Foundation of China (Grant No. E2016202058).