No abstract

This paper describes an improved building treatment approach (IBTA) for use in urban inundation modeling. In this approach, the ground surface elevation was raised by the threshold (h) of the building entrance height to account for both the blockage and storage effect of areas with dense building coverage. A higher roughness coefficient was assigned to the areas where buildings were located to compensate for the resistance effects caused by the inner wall of the structure. The campus of Huazhong University of Science and Technology (HUST) in Wuhan City, China, was used as a case study. Comparison between IBTA and several traditional building treatment approaches suggested that the model results were sensitive to the building treatment method and the threshold used for terrain preprocessing in dense building regions. Furthermore, as the interaction between the surface water flow and dense buildings were adequately represented by using a new terrain preprocessing approach, the proposed IBTA provided better performance in terms of maximum inundation depth and the peak depth time than the traditional approaches in areas with dense building coverage, such as that of the campus.

In this paper, the LSTM model in deep learning is applied to regression analysis, and the LSTM model is used to solve the problems of nonlinearity and data interdependence in regression analysis, so as to improve the traditional regression analysis model. Through the actual modeling application experiment, on the one hand, the prediction accuracy of different model parameters is compared and analyzed, on the other hand, the effectiveness and practicability of LSTM model in multiple regression analysis and prediction are confirmed.

Online course review can objectively reflect the emotional tendency of learners towards the learning effect. This paper proposes a deep neural network based sentiment analysis model for MOOC course reviews. The model uses Bidirectional Long Short-Term Memory Network (BiLSTM) to analyze Chinese semantic. In order to deal with the imbalance of training data set, this paper introduces two methods to balance it and adds dropout mechanism to prevent the over fitting of the model. The model is then applied to the emotional evaluation of MOOC course of “Fundamentals of College Computer Application”. The application results show that the model has achieved good accuracy and can well realize the emotional orientation analysis of online course reviews so as to provide valuable reference for Course Builders.