| 摘要 |
PROBLEM TO BE SOLVED: To solve the problem in which if machine learning is performed based on positive example data and data that is not sure to be a positive example or a negative example, accuracy cannot be estimated that classifies the data into a positive example or a negative example. ! SOLUTION: A classification accuracy estimation device 1 estimates the classification accuracy of a PU classifier 101. The classification accuracy estimation device 1 includes: classification result information acquisition means 103 that acquires classification result information of the PU classifier 101's classification of respective pieces of data included in correct data p and unclear data u that contains correct data p' and incorrect data n; AUC estimation means 105 that calculates AUCP,U from the classification result information to estimate AUC(pcupp',n); accuracy information acquisition means 106 that calculates accuracy information showing a higher value when correct data p' and correct data p are less than inc |
