The dataset is split into 60% training, 20% validation, and 20% testing.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/article>\n
Variables distributions<\/h3>\n
Also, we can calculate the distributions<\/a> for all variables.<\/p>\n The figure below shows a pie chart of the dataset, comparing the proportion of biodegradable (positive) and non-biodegradable (negative) chemicals.<\/p>\n As shown, 33.7% of the samples are biodegradable, while about 66.3% are not.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/article>\n Finally, the input-target correlations<\/a> might indicate which factors most influence biodegradability.<\/p>\n The most correlated variables are SpMax_B(m)<\/b>, SM6_B(m)<\/b>, SpMax_L,<\/b> and SpMax_A<\/b>, whose descriptions are above.<\/p>\n<\/section>\n The second step is to set a neural network<\/a> representing the classification function. For this class of applications, the neural network is composed of:<\/p>\n The scaling layer<\/a> contains the statistics on the inputs calculated from the data file and the method for scaling the input variables.<\/p>\n Here, the minimum and maximum methods have been set. Nevertheless, the mean and standard deviation method would produce very similar results.<\/p>\n For this example, we use a single dense layer<\/a>. This layer has 41 inputs and 1 neuron.<\/p>\n Finally, we set the logistic activation function, as we want the predicted target variable\u00a0to be binary.<\/p>\n<\/section>\n The procedure used to carry out the learning process is called a training strategy<\/a>.<\/p>\n The training strategy is applied to the neural network to obtain the best possible performance.<\/p>\n The type of training is determined by how the parameters in the neural network are adjusted.<\/p>\n This process is composed of two terms:<\/p>\n The loss index<\/a> is the mean squared error<\/a> with L2 regularization<\/a>.\u00a0This is the default loss index for binary classification applications.\u00a0The learning problem can be stated as finding a neural network that minimizes the loss index.<\/p>\n That is a neural network that fits the data set (error term<\/a>) and does not oscillate (regularization term<\/a>).<\/p>\n The optimization algorithm<\/a> that we use is the quasi-Newton method<\/a>.\u00a0This is also the standard optimization algorithm for this type of problem.<\/p>\n The following chart shows how errors decrease with the iterations during training.<\/p>\n The final training and selection errors are training error = 0.108 WSE<\/b> and selection error = 0.105 WSE<\/b>, respectively.<\/p>\n<\/section>\n The objective of the testing analysis<\/a> is to validate the generalization performance of the trained neural network.<\/p>\n To validate a classification technique, we need to compare the values provided by this technique to the observed values.<\/p>\n We can use the ROC curve<\/a> as it is the standard testing method for binary classification projects.<\/p>\n The following table contains the elements of the confusion matrix<\/a>.<\/p>\n This matrix shows the number of negative and positive predicted values versus the real ones.<\/p>\n![](\"https:\/\/www.neuraldesigner.com\/images\/biodegradation-pie-chart.webp\")
<\/p>\nInput-target correlations<\/h3>\n
![](\"https:\/\/www.neuraldesigner.com\/images\/biodegradation-inputs-target-correlation.webp\")
<\/p>\n3. Neural network<\/h2>\n
\n
Scaling layer<\/h3>\n
Dense layer<\/h3>\n
4. Training strategy<\/h2>\n
\n
Loss index<\/h3>\n
Optimization algorithm<\/h3>\n
Training<\/h3>\n
![](\"https:\/\/www.neuraldesigner.com\/images\/biodegradation-quasi-newton.webp\")
<\/p>\n5. Testing analysis<\/h2>\n
ROC curve<\/h3>\n
![](\"https:\/\/www.neuraldesigner.com\/images\/biodegradation-roc.webp\")
<\/p>\nConfusion matrix<\/h3>\n