How to apply machine learning without programming?

Abstract

Now a days, leading companies, like Facebook, Google and Uber has made ML as its central and core part of their operation. For most of the companies, ML has become an important identifier [27]. Machine Learning (ML) lies under Artificial Intelligence (AI) which allows highly accurate predictions from software applications. The system learns, identify patterns and make decisions with minimal interventions from human [27]. ML algorithms predicts output by taking data as input. It means that someone has to build a model which will involve a lot of coding and mathematics. A Google Colab based GUI has been implemented, which will help the non-programmers to make their own models by simple clicks. It’ll involve human interaction but not code. It’ll require the dataset. The user has to select the model or make its own model in case of neural networks. There will be no additional installation of any software. All someone need is a good in ternet connection and basic machine learning and mathematics concepts. The four basic approaches of machine learning are: unsupervised, supervised, reinforcement and deep networks. Supervised learning and deep networks techniques have been implemented which involves regression (single variable and multi-variable regression), artificial neural networks and convolutional neural networks. A user can build his/her desired neural network with desired layers and hyperparameters.