yinbing1030
幼苗
共回答了23个问题采纳率:87% 举报
With the advent of the information age, a lot of data processing has become a focus of research, to find its law and its application. Large amounts of data classification is that people often do, classification data mining is one of the important areas of research. The traditional classification of equilibrium data set classification achieved good results, but the actual data sets are often unbalanced: the sample dataset certain number is far greater than other classes of the number of samples. For the overall classification accuracy based on learning objectives of the traditional classification terms, this imbalance will inevitably lead to too much attention to the majority class classifier sample, so that the minority sample classification performance. In practice, people are more concerned with just a few classes in the dataset, and misclassification costs of those few classes also usually larger than most classes.
Imbalance data can be processed from algorithms and data planes. The starting point of this article is from the algorithm level to classify the data, how to use support vector machine for unbalanced data classification, the main work has the following points:
A: support vector machine theory research.
II: unbalanced data classification research.
Three: given a set of data based on clustering support vector machine division DISVM.
1年前
2