📌Quantity & quality of your data dictate how accurate our model is
📌The outcome of this step is usually a table with some values (features)
📌 If you want to use pre-collected data - get it from sources such as Kaggle or BigQuery Public Datasets
📌Wrangle data and prepare it for training
📌Split data into training, test and evaluation sets
📌Clean data - correct errors, missing values, normalize etc
There are different types of learning to choose from. These 4 are the most common that you usually come across. (will cover these in coming days of #31DaysofML)
📌Goal = make a prediction correctly as often as possible
📌Depends on learning method chosen in step 3
📌Eg: Linear regression algorithm would learn values for weights & biases in y = mx + b
📌Each iteration of process is a training step
📌 Test the model against unseen data (usually 20% of data is set aside for this)
📌 Use some metric or combination of metrics to "measure" objective performance of model