Obviously AI supports several well known Machine Learning algorithms and creates a unique combination of settings (hyper-parameters) to generate custom tailored algorithm for your data.
That's why, on Obviously AI, no two algorithms will ever be the same!
Obviously AI starts by identifying the properties of your data, such as: Number of Rows & Columns, Types of features (Discrete, Continuous, Date/Time, etc), Sparseness, Variance and over 35+ other properties.
Once done, it shortlists top 5 algorithms to use based on the properties of your dataset. Each algorithm has combination of settings called hyper-parameters. Finding the right combination of these settings enables the algorithm make the right predictions.
Obviously AI runs the 5 shortlisted algorithms with thousands of different combinations of settings, each. This means it is running over 10,000+ different algorithm at any given time.
Each algorithm is tested for accuracy and the 1 of the 10,000+ different algorithms, with the highest accuracy is picked and displayed. This entire process happens in less than 1 minute in just one single click!
Below is a list of ML algorithms Obviously AI supports today.