Recently, a lot of Chinese patients consult treatment plans through social networking platforms, but the Chinese medical text contains rich information, including a large number of medical nomenclatures and symptom descriptions. How to build an intelligence model to automatically classify the text information consulted by patients and recommend the correct department for patients is very important. In order to address the problem of insufficient feature extraction from Chinese medical text and low accuracy, this paper proposes a dual channel Chinese medical text classification model. The model extracts feature of Chinese medical text at different granularity, comprehensively and accurately obtains effective feature information, and finally recommends departments for patients according to text classification. One channel of the model focuses on medical nomenclatures, symptoms and other words related to hospital departments, gives different weights, calculates corresponding feature vectors with convolution kernels of different sizes, and then obtains local text representation. The other channel uses the BiGRU network and attention mechanism to obtain text representation, highlighting the important information of the whole sentence, that is, global text representation. Finally, the model uses full connection layer to combine the representation vectors of the two channels, and uses Softmax classifier for classification. The experimental results show that the accuracy, recall and F1-score of the model are improved by 10.65%, 8.94% and 11.62% respectively compared with the baseline models in average, which proves that our model has better performance and robustness.
Tourism recommendation results are affected by many factors. Traditional recommendation methods have problems such as low recommendation accuracy and lack of personalization due to sparse data. This article uses implicit features such as contextual information, time-series travel trajectories, and comment data to address these issues. First, the Long Short-Term Memory (LSTM) network is introduced as the model basis, and deals with the input data of the model such as contextual information, scenic spot information, and tourist comments and so on for feature extraction. Then, the online behavior and long-term interest preference of users are analyzed, using positive feedback and negative feedback mechanism, the Deep Q-Network (DQN) value function of dual-channel mechanism is constructed. Finally, we propose a recommendation strategy, in which, a value evaluation network and a target network are proposed for each agent to learn the optimal strategy. The model is trained on the Yelp, DP, and Tourism datasets covering multiple scenarios to provide users with tourism recommendation services. Compared with baseline models such as Ultra Simplification of Graph Convolutional Networks, DQN, Actor-Critic, and Latent Factor Model, this model has an average increase of 76.61% in accuracy compared with the comparison model, and an average increase of 43.48% in the normalized discounted cumulative gain compared with the baseline model.
15 July 2023 in Information Technology And Control
The number of tourist attractions reviews, travel notes and other texts has grown exponentially in the Internet age. Effectively mining users’ potential opinions and emotions on tourist attractions, and helping to provide users with better recommendation services, which is of great practical significance. This paper proposes a multi-channel neural network model called Pre-BiLSTM combined with a pre-training mechanism. The model uses a combination of coarse and fine- granularity strategies to extract the features of text information such as reviews and travel notes to improve the performance of text sentiment analysis. First, we construct three channels and use the improved BERT and skip-gram methods with negative sampling to vectorize the word-level and vocabulary-level text, respectively, so as to obtain more abundant textual information. Second, we use the pre-training mechanism of BERT to generate deep bidirectional language representation relationships. Third, the vectors of the three channels are input into the BiLSTM network in parallel to extract global and local features. Finally, the model fuses the text features of the three channels and classifies them using SoftMax classifier. Furthermore, numerical experiments are conducted to demonstrate that Pre-BiLSTM outperforms the baselines by 6.27%, 12.83% and 18.12% in average in terms of accuracy, precision and F1-score.
Cooperative navigation of heterogeneous UAV–USV teams in complex island–reef waters involves different action semantics, cross-domain task guidance, and sparse terminal feedback over long horizons. This paper proposes Representation–Intent–Outcome MATD3 (RIO-MATD3), a centralized-training/decentralized-execution framework that assigns these information sources distinct roles in value estimation, policy conditioning, and experience replay. Representation separates centralized value estimation by physical domain and models intra-domain action interactions through agent-specific encoders and attention. Intent conveys UAV task progress to USV policies through guidance moderated by local obstacle risk. Outcome increases the participation of recorded terminal transitions in synchronized replay without goal relabeling. RIO-MATD3 achieves a converged rolling hit rate (CRHR) of 72.57%. Across three independent test conditions, the joint success rates are 70.8%, 71.4%, and 58.6%, respectively. The results indicate complementary roles for domain-specific value estimation, risk-conditioned guidance, and terminal-outcome reuse.