Coding Activities by Bhargav Venkat
Try Different Layers
Open the link to the page given below. Consider you have been given a dataset - Pima Indians onset of diabetes dataset. It describes patient medical record data for Pima Indians and whether they had an onset of diabetes within five years. As such, it is a binary classification problem (onset of diabetes as one or not as 0). This dataset consists of a total of 3 features/columns, including diabetes onset. Now, set the training to testing ratio to 80%. Now try adding hidden layers and neurons in each hidden layer and create a google document to enter values of testing loss for the following entries - 1) 1 Hidden layer (3 neurons) 2) 2 Hidden layers (3 neurons) 3) 3 Hidden layers (3 neurons) 4) 4 Hidden layers (3 neurons) 5) 3 Hidden layers with eight neurons 6) 3 Hidden layers with six neurons
Quick Draw
Play this game to understand how AI can recognize the word through doodles drawn by the player. In Quick Draw, a model has been created which has been trained on multiple images of different objects, and it gets trained through doodles drawn by players as well. And this is how AI is making this game smarter every time.
Supervised vs. Unsupervised Algorithms
Let’s do this activity to summarise the concept of Supervised and Unsupervised ML algorithms. Write the names of Supervised and Unsupervised ML algorithms learned until now with some applications using your learnings. You can use some graphs and pictures for a better understanding. Go to file -> Make a copy -> Rename it -> Start the solution -> Save it
DBSCAN & Hierarchical Clustering for Customer
Do this activity to see how DBSCAN and Hierarchical clustering works. The input data set contains information (ID, age, gender, income, spending score) about the mall's customers. A spending Score is something we assign to the customer based on our defined parameters like the customer’s behavior and purchasing data. Drop the customer ID column, as it is not correlated with income. Apply the DBSCAN and Hierarchical clustering to the customer dataset.
K-Means Clustering for Customer
Do this activity to see how K-Means clustering works. The input data set contains information (ID, age, gender, income, spending score) about the mall's customers. A spending Score is something we assign to the customer based on our defined parameters like the customer’s behavior and purchasing data. Apply the K-Means algorithm to the customer dataset.
Supervised vs. Unsupervised
Through this activity, let’s recall the supervised machine learning algorithms and see the difference between supervised and unsupervised ML algorithms. We have given the dataset of Housing in the USA. Apply linear regression and K Means Clustering to predict the price of a house in the USA using parameters like Avg. Area income, house age, number of rooms, area population, and address. K Means is a clustering algorithm that we will learn in other lessons. Still, for now, you can take the reference from the solution of the activity to see the actual difference between the supervised and unsupervised machine learning algorithms.