Li Qiaoqin

TL;DR: In this paper, a neural network CNN is used to carry out feature extraction on Chinese character strokes; the Chinese character feature vector of each character and the corresponding stroke feature vector are connected in series and input into the bidirectional long-short-term memory model, and the same points of the Chinese characters can be learned by using the strokes of Chinese characters through the neural network convolutional neural network.

TL;DR: Wang et al. as mentioned in this paper presented a method and system for identifying traditional Chinese medicine pharmacological actions based on machine learning and text rules, which consists of constructing a training corpus based on a BIO rule, extracting and digitizing text features, and constructing a pharmacological action identification model by using multi-classification SVM; and finally, post-processing an annotation result output by the SVM model using a rule-based error-driven learning(TBL) method to improve the accuracy of entity recognition.

TL;DR: Wang et al. as mentioned in this paper proposed a deep learning-based method for the automatic recognition of the urine color of the Tibetan medicine, where the features are automatically learned from training data through the deep convolutional neural network instead of manually designed features.

TL;DR: In this paper, a traditional Chinese medicine intelligent syndrome differentiation auxiliary decision-making method based on the topic model technology was proposed, which comprises the following steps: normalizing symptom names in a medical record set; preprocessing a medical case data set; performing word segmentation processing on each medical case by using a language technology platform tool; generating a medical cases topic model to obtain all topics implied in the medical case set; constructing a standard syndrome database based on Internal Science of the Traditional Chinese Medicine, and obtaining the labels of the topics by calculating the similarity between symptom groups under the themes

TL;DR: In this article, a human body falling detection method based on mobile sensor combined equipment is presented, which comprises the following steps: acquiring human body user sensor data based on a wearable sensor system; and performing numerical normalization processing on the acquired sensor data; classifying the acquired sensors data by adopting a CorrRNN model based on time sequence multi-modal learning; collecting data through waist and wrist sensors to construct classifiers respectively, and weighing and combining classification results to obtain a fall category judgment result.