Incremental learning of LDA model for Chinese writer adaptation

Study on Influencing Factors of Construction Workers' Unsafe Behavior Based on Text Mining

The unsafe behavior of construction workers is the key cause of safety accidents. The accident investigation report contains rich experience and lessons, which can be used to prevent and reduce the occurrence of safety accidents. In order to draw lessons from the accident and realize knowledge sharing and reuse, this paper uses text mining technology to analyze the data of 500 construction accident investigation reports in Shenzhen, China. Firstly, a Latent Dirichlet Allocation (LDA) topic model is used to identify the unsafe behavior of construction workers and its influencing factors. Then, with the help of Social Network Analysis, the importance of influencing factors and the relationship between them are identified. The results show that weak safety awareness, operating regulations, supervision dereliction of duty, equipment resources, and inadequate supervision of the construction party are the key and important factors. It is also found that there are correlations between weak safety awareness and supervision dereliction of duty, between equipment resources and poor construction environment, between organization and coordination and inadequate supervision of the construction party, and between operating regulations and hidden dangers investigation. This study not only helps to improve the theoretical system in the field of construction workers’ unsafe behavior but also helps managers to find the key control direction of construction safety, so as to effectively curb unsafe behavior of construction workers and improve the level of safety management.

Writer adaptation with style transfer mapping

Adapting a writer-independent classifier toward the unique handwriting style of a particular writer has the potential to significantly increase accuracy for personalized handwriting recognition. This paper proposes a novel framework of style transfer mapping (STM) for writer adaptation. The STM is a writer-specific class-independent feature transformation which has a closed-form solution. After style transfer mapping, the data of different writers are projected onto a style-free space, where the writer-independent classifier needs no change to classify the transformed data and can achieve significantly higher accuracy. The framework of STM can be combined with different types of classifiers for supervised, unsupervised, and semi-supervised adaptation, where writer-specific data can be either labeled or unlabeled and need not cover all classes. In this paper, we combine STM with the state-of-the-art classifiers for large-category Chinese handwriting recognition: learning vector quantization (LVQ) and modified quadratic discriminant function (MQDF). Experiments on the online Chinese handwriting database CASIA-OLHWDB demonstrate that STM-based adaptation is very efficient and effective in improving classification accuracy. Semi-supervised adaptation achieves the best performance, while unsupervised adaptation is even better than supervised adaptation. On handwritten text data, semi-supervised adaptation achieves error reduction rates 31.95 and 25.00 percent by LVQ and MQDF, respectively.