People and post matching method based on symmetric comparative learning

The invention discloses a person and post matching method based on symmetric comparative learning, and belongs to the field of computer data processing. According to the method, resume texts are divided into regular texts and irregular texts, strategies are adopted for the regular texts to determine remaining resumes, the irregular texts are classified according to work experiences, project experiences, post requirements and job requirements, and any part is subjected to unsupervised text semantic representation; a key loss calculation strategy adopts a symmetric contrast learning function, the defect that a traditional loss function cannot completely meet the contrast learning thought is overcome, and the distance between a real sample and an enhanced sample can be calculated on the whole. And finally, calculating a semantic matching degree between irregular texts between resume information such as work experiences and recruitment information according to the calculated text semantic representation through a