GetRichWithDiamonds

Inspiration

Have you ever asked yourself, how are diamonds priced? Well, this article talks about the diamonds price prediction based on their cut, colour, clarity & other attributes and it also covers the building a simple linear regression model.

Diamond prices can vary hugely depending on a diamond’s shape, cut quality, clarity and color. For example, the cost of a one carat diamond can range from just $1,500 to more than $16,000 for an extremely well cut, high quality diamond, while a two carat diamond could cost as little as $6,000 or as much as $80,000 based on its shape, cut, clarity and color grades.

What it does

This project predicts the Price of diamonds using Machine learning algorithms like Linear Regression, RandomForestRegressor. It compares different paramters like cut quality, clarity, color and carat weight of diamonds.

How we built it

1: Load the dataset

2: Perform the exploratory data analysis (EDA)

3: Prepare the dataset for training

4: Create a linear regression model

5: Train the model to fit the data

6: Make predictions using the trained model

7: Create Random forest regressor

8: Perform hyperparamater tuning

9: Make predictions and calculate accuracy

Challenges we ran into

As you can see, this is a huge range, with some diamonds costing as much as 10 times higher than other diamonds of the same carat weight.

Diamond prices depend on such a wide range of factors that’s very difficult to give an accurate price estimate for “diamonds” as a whole. The biggest of these are the four Cs, which we briefly mentioned earlier — cut quality, clarity, color and carat weight.

Accomplishments that we're proud of

Using Linear Regression model we got an accuracy of almost 70% but using RandomForestRegressor we have successfully achieved an accuracy of almost 91%.