本文针对传统机器学习算法需要大量标注数据，数据分布差异大，以及需要大量

This article requires a lot of annotation data for the traditional machine learning algorithms, data distribution big difference, and requires a lot of prior information and other issues, we proposed a frequency-hopping signal modulation depth migration methods to identify learning and multi-classification SVM.

In this paper, a method of identifying the frequency-hopping signal modulation based on deep migration learning and multi-classification SVM is proposed for the traditional machine learning algorithms, which require a large amount of labeling data, the large difference in data distribution, and the need for a large amount of a priorinformation.

Aiming at the problems of traditional machine learning algorithm, such as large amount of labeled data, large difference of data distribution, and a large amount of prior information, this paper proposes a modulation recognition method of FH signal based on deep transfer learning and multi classification SVM.<br>