Deep Forest-Based Intelligent Yield Predicting of Buchwald-Hartwig Coupling Reaction

Abstract

Buchwald-Hartwig coupling reaction is widely used in organic chemical synthesis, yield prediction is particularly important. In 2018, Science reported a yield prediction method based on random forest, but this method lacks feature learning. Therefore, an intelligent prediction and analysis method of coupling reaction yield based on deep forest is proposed. Combined with the advantages of deep learning and ensemble learning, the new deep model in the form of non-neural network is explored, which has good characterization learning ability and low difficulty in adjusting parameters, realizes the efficient prediction of chemical reaction, and analyzes the factors that have a significant impact on the prediction of reaction yield.

ChemCNet: An Explainable Integrated Model for Intelligent Analyzing Chemistry Synthesis Reactions

“…With theoretical modeling and technological innovations, complex chemical reactions can be simulated, synthesized with the high-speed processing power of computers, and the cross-fertilization of artificial intelligence algorithms with chemical disciplines is of great importance to advance academic research [6]. In 2018, D. T. Ahneman et al [7] reported the prediction of the yield of the Buchwald-Hartwig amination reaction by random forests, an advanced study of machine learning methods in the field of multidimensional chemical space prediction; M. H. S. Segler et al [8] proposed the use of recurrent neural networks as a generative model for molecular structures; J. Dong et al [9] used the XGBoost model as a prediction model, and X. H. Mu et al [10] used Deep Forest as a model, both of which improved the prediction accuracy. However, these works are not strong enough for deep feature mining of data, and further feature learning is a direction worth thinking about.…”

OCS-TGBM: Intelligent Analysis of Organic Chemical Synthesis Based on Topological Data Analysis and LightGBM

“…In recent years, machine learning (ML) as an efficient method has been gradually applied in the field of bioinformatics and chemistry [10,11]. It shows more and more competitiveness in the research of chemical reaction prediction [12][13][14], drug performance prediction [15][16][17][18][19], screening for target compounds [20][21][22][23], molecular material design [24][25][26]. Recently, researchers considered using ensemble tree models to predict the performance of chemical reactions, which are easy to analyze and interpret.…”

Regression Prediction of Coupling Reaction Yield Based on Attention-Driven Convolutional Neural Network

“…However, this method is point prediction based on feature descriptors and lacks feature learning. Based on this, our team previously proposed intelligent yield prediction based on quantile regression forest and deep forest respectively [11,12].…”