Aiming at the problem of ensuring the effectiveness and mining analysis value of trajectory semantic data while realizing personalized privacy protection of vehicle trajectory data， a vehicle trajectory semantic protection mechanism based on improved Generative Adversarial Network （GAN） was proposed. In this mechanism： firstly， a position sensitivity grading and semantic annotation method based on Hidden Markov Model （HMM） was designed to extract the effective stop points from vehicle trajectories， and then the stop points were divided into different sensitive levels and annotated semantically. Secondly， Long Short-Term Memory （LSTM） network was introduced into the improved GAN to construct the semantic trajectory model based on the dynamic GAN， and the GAN model was used for training to generate high-quality synthetic trajectories. Finally， for the stop points in synthetic trajectories that required further privacy protection， a differential privacy personalized protection algorithm combining the position sensitivity levels was proposed， which assigned privacy budgets to the stop points according to their sensitivity level and correlation between the stop points， and noise was injected by combining with the Laplace mechanism to achieve the privacy protection， so as to maximize the usability of the trajectory data after protection. Experimental results show that compared to the LSTM-TrajGAN model， the proposed mechanism reduces the Mutual Information （MI） value by 27.58% and improves the semantic trajectory similarity by 24.4%. It can be seen that the proposed mechanism protects user privacy effectively while ensuring the usability of semantic trajectory data.

In order to meet the needs of modern society for weather forecast and early warning service, a real-time weather forecast and abnormal weather early warning system was researched and implemented in the Android system. Based on the decision tree algorithm of C4.5 algorithm, the warning classification problem was resolved. By means of extracting the attributes with maximum gain rate as the features of training sample, a decision tree was built. A model of decision tree was got by the pruning weather warning evaluation and analysis and application were made on this model. The experimental results show that this method has advantages in the assessment of classification accuracy, with correct classification rate up to 85.8%.

The microscopic simulation model based on artificial potential field and Agent was proposed from the individual perspective to evacuate in fire environment. First, various factors of fire, doors, obstacles etc which can affect peoples’ behavior were quantified to construct artificial potential field system. Second, the individual was as object and movement direction can be determined through stress analysis. Finally after whether to open the exits and whether to involve personnel persuasion were compared, crowd evacuation was simulated reasonably. The evacuation system is realized based on Visual C++ and can be authentic to simulate the crowd evacuation process from the experimental results.

This paper proposed a method of extracting sentiment topic sentences. Firstly semantic concepts of a text were evaluated in order to determine which concepts were related to the topic of a text. And the concepts related to the topic were regarded as topic concepts. Sentences including one or more topic concepts were defined as candidate sentences. Significance of every candidate sentences was calculated in order to which ones were topic sentences in the text. Conditional random field model was adopted and two kinds of feature were used in the model training, and one feature was polarity of sentiment and the other feature was transferring words. This approach excluded sentences that were not related to the topic of the text, and eliminated the influence brought by these sentences. Therefore, precision of sentiment analysis is effectively improved.