Aimed at providing efficient training data for neural translation quality estimation model, a pseudo data construction method for target dataset is proposed, the model is trained by two stage model training method: pre training based on pseudo data and fine tuning. The experimental design of different pseudo data scale is carried out. The experiment results show that the machine translation quality estimation model trained by the pseudo data has significantly improved in the correlation between the scores given by human and the artificial scores.
This study targets on the practical process of travel survey in Qianmen, Beijing and examines the problems derived from the survey. Their characteristics and the reasons of being generated are stated. The paper focuses on survey organization and its institutional obstacles, the survey design, survey sampling techniques, the choice and training of surveyors and the survey timing. Based on the theoretical researches, the advices towards the innovation of travel survey methods are proposed.