Urban land simulation method based on Distance-SMOTE-CNN-CA model

Abstract

The invention relates to an urban land simulation method based on a Distance-SMOTE-CNN-CA model, and belongs to the field of urban dynamic simulation, and the method comprises the steps: collecting multi-stage land utilization data and driving factor data in a research region; performing stratified sampling on the data set; balancing Distance-SMOTE training data to different magnitudes, and selecting better balanced data by comparison; equalizing the training data to an optimal data magnitude by using a Distance-SMOTE equalization algorithm; performing land utilization data prediction trainingin the CNN-CA model; inputting the first-stage test set into the CNN-CA model to obtain a second-stage land utilization data prediction result; performing Kappa index test on the second-stage land utilization data prediction result and the second-stage actual land utilization data to check the overall precision and consistency; and inputting the test set of the current year into the CNN-CA model,and predicting future land utilization data.

Description

technical field [0001] The invention belongs to the field of urban dynamic simulation and relates to an urban land use simulation method based on the Distance-SMOTE-CNN-CA model. Background technique [0002] In the field of machine learning, data imbalance is a hot issue. Data imbalance refers to the large difference in the number of samples of the majority class and the number of samples of the minority class in the sample set. Although the overall accuracy of the model is high, there are cases where the accuracy of the minority class samples is relatively low. Minority class samples are also data Focus on one important item. Therefore, under the premise of ensuring the overall accuracy, how to improve the accuracy of minority class samples has become a problem. In the face of data imbalance, it can be solved at two levels: data preprocessing level and algorithm level. Data preprocessing is mainly sampling, and oversampling or undersampling methods are usually used to p...

AI Technical Summary

Problems solved by technology

Data imbalance refers to the large difference in the number of samples of the majority class and the number of samples of the minority class in the sample set. Although the overall accuracy of the model is high, there are cases where the accuracy of the minority class samples is relatively low. Minority class samples are also data Concentrate on an important item

Images