A neural network model was created to predict toxicity of chemicals. The model was trained on 282 chemicals and tested on 286. It performed well, with an accuracy of 83% overall. The model can also create maps to show how the chemicals are grouped together and identify any unusual ones.