数字分类习题

Source Codes of Exercises in CMSC 170 on Digit Classification for Machine-generated and single and multiple handwritten digits by Daryl Ancheta and Paul Andrew Manuales, UPLB BS Computer Science students of Sir Vladimir Mariano. The following exercises used a total of 200 training samples(20 for each digit 0-9) for machine, single and multiple. The algorithm and features used were tested using 50 random samples for each category. The results are available at http://cmsc170.blogspot.com Nearest Mean Classifier Exercises: Exercise # 3 - NEAREST MEAN CLASSIFIER Exercise # 4 - DECISION REGIONS AND DECISION BOUNDARIES Exercise # 5 - PERFORMANCE EVALUATION k-Nearest Neighbor Classifier Exercises: Exercise # 6 - K-NEAREST NEIGHBOR CLASSIFIER Exercise # 7 - K-N-N DECISION REGIONS