Week 2: Regressions

Ishan B -

When making a choice of where to open a business, there is no correct answer. Each location has their own pros and cons, and the business owners preferences also factor into the decision. Because of this, I plan to use a regression model as that can help rank the different locations rather than just outputting a single location that the model thinks is best. Last week, my research on regressions was mostly for what they are, how they have been used before, and what are the pros and cons of each. This week, I have been researching more on the side of how they work in relation to machine learning.

Linear Regression

Linear regression models work by making their prediction using a weighted sum of the input features of the data and adding on a bias term/ intercept term to it. The bias term is just a number that helps shift predictions of the model up or down. Linear regression is a possibility for my use case as each business owner has certain factors they value more than others, which partially works like a weighted sum in itself as they are assigning their own weights to the factors.

The equation is: y = θ0 + θ1x1 + θ2x2 + … + θnxn

Polynomial Regression

Polynomial regression models work the same way as linear regression models except that we are adding powers to every term. Polynomial regression is also a possibility for my use case as the rankings of each location may not be perfectly linear.

Ridge Regression and Lasso Regression

Both Ridge and Lasso Regression are similar to linear regression except that we are adding a regularization term to the cost function. This means that ridge and lasso regression can be used to help tune linear regression models, potentially boosting the accuracy of the model.

Logistic Regression

Logistic Regression is a regression model that is used for classification, which is not the use case of my model so I will not be using it.

More Posts

Comments:

All viewpoints are welcome but profane, threatening, disrespectful, or harassing comments will not be tolerated and are subject to moderation up to, and including, full deletion.

    shreyash_p
    Great work, Ishan! I love how you’re breaking down the different regression models and considering their applications for your business location decision-making process. It sounds like a really thoughtful approach. How do you plan to weigh the various factors for each location? Will you base that on data, or will you incorporate business owners' personal preferences more heavily?
    February 20, 2025 at 2:16 pm - Reply
    Anshul Baddi
    Hey Ishan, I like how you have broken down the different kinds of regression models. I was wondering if you would consider a large language model because it would let the client manually input all the factors they need.
    February 21, 2025 at 10:29 am - Reply
    ishan_b
    Thanks for the question Shreyash! Initially, for different types of businesses (education, technology, etc.) I will set base weights based on data. After I have done that, I will create a system where the business owner ranks which factors they care about most and shift the weights based on those rankings.
    March 2, 2025 at 11:33 am - Reply
    ishan_b
    Thanks for the question Anshul! One of the main reason that I chose to use regression models instead of different models is the capability to easily hypertune the model after initial training. A large language model would definitely allow for the business owners to input more factors than the regression model will allow, however I wanted to prioritize being able to hypertune the model.
    March 2, 2025 at 11:35 am - Reply

Leave a Reply

Your email address will not be published. Required fields are marked *