Random Case Analysis of Inductive Learning Algorithms

摘要: In machine learning, it is important to reduce computational time analyze learning algorithms. Some researchers have attempted understand algorithms by experimenting them on a variety of domains. Others presented theoretical methods algorithm using approximately mathematical model. The model has some deficiency that, if the too simplified, may lose essential behavior original algorithm. Furthermore, experimental analyses are based only informal task, whereas address worst case. Therefore, results quite different from empirical results. our framework, called random case analysis, we adopt idea randomized By can predict various aspects algorithm's behavior, and require less than other analyses. easily apply framework practical