Tutorial: Using machine learning to train AI bot for data analysis automation

Choose the Regression model if you want to train your AI bot to predict a value for some input data. The inputs can be continuous our discrete and the output is a continuous value. A first example of regression jobs is an AI bot that predict residential energy consumption based on the building size and neighborhood energy usage history.

To load features from a sample, proceed as follow: Go the FEATURE TAB, Choose you model and data type and then validate. Next, upload one of your sample file (or the whole sample file if you selected Single CSV as data type) and click on the Validate button. If you selected CSV as data type, the found features will show up at the bottom of the page.

Classifier bots analyze data from sample and maps these input unto the relevant label among given categories. In other word, the label is the value which is meant to be predicted (output) by the AI neural network, given the sample’s specifications (features).

The number of required layer depends mainly on the complexity of the learning task. While a single hidden layer is enough for most models, using a certain number of abstraction layers is usually needed for problems such as the image classification.

Choosing the optimal training parameters improves your AI bot performance at output predicting and reduce the computational time. Whichever your neural network architecture, the following parameters are required:

By selecting the proper regularization type and increasing the regularization rate, you can force your neural network to remain simple and then avoiding point-to-point mapping. You should note that while regularization and dropout helps preventing overfitting, it can also penalize the learning process and reduce the prediction accuracy.