A Novel Multi-Source Domain Adaptation Method with Dempster–Shafer Evidence Theory for Cross-Domain Classification

In this era of big data, Multi-source Domain Adaptation (MDA) becomes more and popular is employed to make full use available source data collected from several different, but related domains. Although multiple domains provide much information, the processing domain shifts challenging, especially in learning a common domain-invariant representation for all Moreover, it counter-intuitive treat equally as most existing MDA algorithms do. Therefore, domain-specific distribution each source–target pair aligned, respectively. Nevertheless, hard combine adaptation outputs different classifiers effectively, because ambiguity on category boundary. Subjective Logic (SL) introduced measure uncertainty (credibility) classifier, so that could be bridged with DST. Due advantage information fusion, Dempster–Shafer evidence Theory (DST) utilized reduce boundary output reasonable decisions by combining based uncertainty. Finally, extensive comparative experiments three benchmark datasets cross-domain image classification are conducted evaluate performance proposed method via various aspects.