부산대학교 도서관

Climate change is surely one of the major concerns in the 21st century. Various research efforts to understand the future of the world's climate are being made. The climate projections based on standard rational-based physics and global system models are well known for this. In this study, we examine the effectiveness of entire phenomenological model which uses top-down approach to obtain 21 century projections of precipitation in the country of India using dimensionality reduction technique and LSTM, GRU and big data of global general circulation models (GCMs). As part of research work, I referred different literatures related to same. To train this neural network faster technique named UMAP (Uniform Manifold Approximation Projection) is applied here for getting better result.