Dry Weight Prediction of Wedelia trilobata and Wedelia chinensis by Using Artificial Neural Network and MultipleLinear Regression Models

In China, Wedelia trilobata (WT) is among the top most invasive plant species. The prediction of its growth, using different efficient methods under environmental conditions, optimal objective ecological research. For this purpose, and native species chinensis (WC) were grown in mixed cultures levels submergence eutrophication. multiple linear regression (MLR) artificial neural network (ANN) models constructed, with morphological traits as input order to predict dry weight output for both Correlation stepwise analysis (SWR) used find best variables ANN MLR models. Plant height, number nodes, chlorophyll content, leaf nitrogen, leaves, photosynthesis, stomatal conductance WC. same WT, addition root length. A Levenberg–Marquart learning algorithm, back propagation training Sigmoid Axon transfer function, one hidden layer, four six neurons WC respectively, was created. model (7-4-1) has a coefficient determination (R2) 0.98, mean square error (RMSE) 0.003, absolute (MAE) 0.001. On other hand, WT (8-6-1) R2 RMSE 0.018, MAE 0.004. According errors values, more accurate than one. sensitivity analysis, height nodes are important that support growth eutrophication conditions. This study provides us new method control species’ spread habitats.