Presentation | 2012-08-01 A Study on Grading Error Detection of Entrance Examination based on the Pattern Categorization Using Asymmetric Partial Normal Distribution Masato SUZUKI, Daisuke KITAKOSHI, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | A wrong answer detection algorithm with sufficient accuracy is required in the grading support system of entrance examination. A pattern recognition algorithm is effective for distinguishing a certain class from other classes, but it is premised that features extracted from character picture is following a normal distribution, in many cases. Therefore, it is difficult to assume a normal distribution as a distribution of characteristic features, because the distribution of them is distorted when wrong answers are contained. In this manuscript, we propose the asymmetric partial normal distribution, which made the normal distribution distorted. We can estimate the population distribution from samples, since the statistics value up to the third order can be solved analytically, in the asymmetric partial normal distribution. So, we can detect wrong answers successively with hypothesis testing based on the estimated population distribution. In experiments, we can detect 95.8% of the whole wrong answer using our algorithm, and it is found that our algorithm is useful to detect wrong answers for grading support system of entrance examination. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Asymmetric partial normal distribution / Grading support system of entrance examination / Error detection |
Paper # | DE2012-17 |
Date of Issue |
Conference Information | |
Committee | DE |
---|---|
Conference Date | 2012/7/25(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Data Engineering (DE) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | A Study on Grading Error Detection of Entrance Examination based on the Pattern Categorization Using Asymmetric Partial Normal Distribution |
Sub Title (in English) | |
Keyword(1) | Asymmetric partial normal distribution |
Keyword(2) | Grading support system of entrance examination |
Keyword(3) | Error detection |
1st Author's Name | Masato SUZUKI |
1st Author's Affiliation | Department of Computer Science, Tokyo National College of Technology() |
2nd Author's Name | Daisuke KITAKOSHI |
2nd Author's Affiliation | Department of Computer Science, Tokyo National College of Technology |
Date | 2012-08-01 |
Paper # | DE2012-17 |
Volume (vol) | vol.112 |
Number (no) | 172 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |