Predicting Acute Aquatic Toxicity Towards Fathead Minnow (Pimephales Promelas) Using Neuro-Fuzzy Inference System (ANFIS)

The variety of chemicals used in everyday life tend to have a significant impact on the environment, only one of which is the negative impact on the earth's bodies of water and its inhabitants. This paper aims to predict the acute aquatic toxicity rate of various chemicals towards the flathead minnow using a neuro-fuzzy approach given only six different molecular descriptors. Actual data parameters from a previously conducted research project on quantitative structure-activity relationship (QSAR) prediction models will be utilized as the training and testing data for the network. In testing the data, comparisons will be made between the various fuzzy inference system (FIS) models and their respective performances. Likewise, the generated fuzzy rules will be analyzed and assessed using a set of testing data to check for accuracy. Results show both training and testing errors to be at acceptable levels, thus, proving the feasibility of determining acute aquatic toxicity using adaptive neuro-fuzzy inference system (ANFIS) models.