Traditional text representation techniques will produce high-dimensional and sparse vectors for short texts since they have fewer words, which increases the likelihood of noise interference and lowers the accuracy of short text clustering. An alternative model, CASA-Bert-K, is investigated to overcome the aforementioned issues. First, In order to add contextual features into the text representation, context-awareness self-attention networks are combined with Bert. Secondly, To de-optimize the representation of short texts, a simple and effective data augmentation approach to contrast learning is introduced. The text representation module and the clustering module are jointly trained to optimize both text representation and clustering in order to further improve the clustering accuracy. The experimental findings show the researched model's effectiveness in raising the accuracy rate through the use of three datasets in the tests.
In order to effectively monitor and detect the characteristics and abnormal behavior of network traffic, this paper uses convolution neural network model to realize the classification and prediction of network data. Firstly, a system model including data preprocessing module, data hierarchy module and convolution neural network module is established. Secondly, the data is preprocessed and normalized by convolution operation. Then, the pooled pool is used to select features, and the back-propagation algorithm and gradient descent algorithm are used to optimize the model to complete feature extraction and training. Based on this, the characteristics and abnormal behavior of network traffic can be monitored and detected. In the experiment, the KDDCUP 1999 data set and UNSW-NB15 data set were used for system testing and compared with various methods. The experimental results show that the detection accuracy, precision, recall and F1 index of the proposed system are more than 10% higher than those of other methods, showing good performance and high application value.
Using wireless body area networks to monitor and recognize human motion is one of the research hotspots in wireless body area networks, and gesture is an important component of human motion. Gesture, as a natural, intuitive, and easy to learn human-computer interaction method, can provide operational convenience for health monitoring in wireless body area networks. Gestures are divided into static and dynamic gestures. Static gestures rely solely on the external shape and contour information of the hand to convey information. Dynamic gestures also include gesture timing information for gesture movements, which can express richer and more accurate information. Data acquisition methods for gesture movements include data glove based, vision based, EMG based, and inertial sensor based methods. This paper studies a dynamic gesture recognition method based on acceleration sensors.
The research results of Chinese Medical academic articles are more accurate and representative than other medical texts. This paper studies the extraction method of medical entity recognition in Chinese academic articles on hypertension. This paper proposes a BiLSTM-CRF (Bidirectional Long Short-Term Memory and Conditional Random Fields) model blended with TFE (Text Feature Embedding) and BERT (Bidirectional Encoder Representation from Transformers) for Chinese medical entity recognition. BERT model was used for text pre-processing while text features of Chinese medical academic articles were obtained, and then the annotation and classification of medical entity were completed via three models. Compared with the BiLSTM-CRF model and BERT-BiLSTM-CRF model, the recognition effect of the BERT-TFE-BiLSTM-CRF model proposed in this study was better than the other two models and the F1 value of model was 6.9% and 3.3% higher respectively. The BERT-TFE-BiLSTM-CRF model can effectively recognize medical entity in Chinese academic articles on hypertension.
With the trend of globalization, modern information technology breathes new life into all walks of life. Concerning the complexity of English phonetic training, it’s urgent to explore a practical system related to information technology. Cloud service system provides a possible new educational technique, which would help English phonetic training achieve man-machine interaction, computing, Cloud storing, and integrating conveniently and efficiently. This paper introduces the background of English phonetic training under the circumstances of Cloud Platform. Then, current information technologies related to English phonetic training are exhibited. It proposes the feasible ways to construct English phonetic training system on Cloud Platform. By experiment, it makes quantitative and qualitative analysis, and further testifies the validity of this construction by analysing and discussing the results of the experiment due to statistical data. It draws the conclusion that to construct English phonetic training system on Cloud Platform is feasible and effective. It helps arouse more scholars to more concern about application of English phonetic training system on Cloud Platform.
