Aniket Sharma

Machine Learning Zoomcamp Update: Tuesday, 14 September 2021

Today, I completed the Sessions 2.1, 2.2, 2.3, 2.4 and 2.5.

Session 2.1 - Car price prediction project

In this session Alexey Grigorev introduced this week's lessions and the project we will be working on this week.

Key takeaways:

• We will learn regression by working on a project using Car Features and MSRP dataset on Kaggle.
• The Project plan is as follows:
  • Prepare data and do EDA (Exploratory Data Analysis)
  • Use linear regression for predicting price
  • Understanding the internals of linear regression
  • Evaluating the model using RMSE
  • Feature engineering
  • Regularization
  • Using the model

Session 2.2 - Data preparation

This session explains how to perform data cleaning and prepare data for training.

Key takeaways:

• Data should be consistent.

Session 2.3 - Exploratory data analysis

This session explains Exploratory data analysis.

Key takeaways:

• Before training we should try to find any patterns if there are in the data.
• Using np.log1p method we can remove the tail if there is in data distribution.

Session 2.4 - Setting up the validation framework

This session explains train, validation and test split.

Key takeaways:

• The data should be first shuffled before creating the splits.
• We should set the random seed to make reproducible results.

Session 2.5 - Linear regression

This session gave an introduction to linear regression.

Key takeaways:

• We can use the np.expm1 to convert model result to actual predictions if we have used np.log1p previously.
• Prediction in linear regression: g(x_i) = w₀ + ∑_j=1ⁿ w_jx_ij, where there are n features

Estimated Time Taken: 1 hour 10 minutes