Emulation of dynamic simulation models of vegetation growth
- Most of the plant growth simulation has ultra-high time and space consumption.
- A surrogate model is a good replacement for those high consumption simulators.
- It is a challenge to design a surrogate model so that it performs as well as the original scientific model while remaining computationally efficient remains.
- Short time-stamp prediction of NPP.
- Simulated data (NPP)’s whole time-series emulation process with multiple period prediction.
The data is from the JeDi-DGVM which is provided by the project manager and supervisor. Because of the security issue all data from JeDi-DGVM will not be include in the repository, only the result will be displayed
- ARIMA_models: ARIMA models which are applies by the cluster.
- ARIMA_summary: Summaries of ARIMA models which includes the key parameters, estimator, standard error and p-value.
- cluster_models: main code which includes the process of data preprocessing, uilts, and models of cluster, ARIMA and LSTM.
- LSTM_models: LSTM models which are applies by the cluster.
- LSTM_summary:Summaries of LSTM models which includes the information of each layer.
- Random_Forest: Whole process for Random_forest with data preprocessing and modeling.
- numpy: Arrays or label which is for modeling or plotting.
- plot: Plots for summarizing and evaluation
Professor In Climate Science
School of Geography, Earth and Atmospheric Sciences of University of Melbourne
Research Fellow In Data Science
Mathematics and Statistics of University of Melbourne
- Yuze qu:[email protected]
- Zhili Chen (Emile):[email protected]
- Zening Zhang (Nathan): [email protected]
- Ziqian Wu (Randy):[email protected]
- Jiachen Huo (Cassiel):[email protected]
- Canva: https://www.canva.cn/design/DAFBCMwo31g/1Fq1tvp-h1svuryfYaXv3A/view?utm_content=DAFBCMwo31g&utm_campaign=designshare&utm_medium=link2&utm_source=sharebutton
- Google doc: https://docs.google.com/presentation/d/1c67JZ0hg1SJYNk6pikyPvD__esQDJoaduzWXMD8hYZM/edit?usp=sharing
- Youtube: https://www.youtube.com/watch?v=0DoxLSOES4Q
JeDi-DGVM, Surrogate Model, Random Forest, Clustering, ARIMA, Long Short-Term Memory