hsc012
commented
2 years ago Hi Stephen, thank you for the feedback. We just noticed the 0.75/1 for Metrics but did not see any comments regarding why we got points off so we were wondering if this was a mistake, and if not, what should we change? Thank you!
Total Points: 7/8
Title & Abstract (0.5/1) The language you use to describe your approach is incredibly vague and the tone is too informal. Try to be more precise in your language when defining an Abstract.
Something like: "Our goal is train a regression model with a large dataset of app usage (or some data with lots of features) and ratings data from the Google Play Store to predict ratings for apps."
Then you should explain briefly why this is important. What's the purpose of this?
You also say things like "we were lucky enough to find data on , and with this our goal is to predict ." You should state your approach concisely and clearly.
Also you should be using more than one model. For this class, a simple OLS / linear regression is not enough for the scope of the project!
I'd suggest adding other regression models, maybe some that you mention in your Background section.
Background (1/1) Nice!
Problem Statement (0.75/1) Don't group these ML "solutions" together; Building a linear regression/OLS model would entail one-hot-encoding, data wrangling, and other Python functions. Those data cleaning/processing steps are not independent solutions, but methods to prepare data for the ML model. It can be thought of as either 1 encompassing solution, or you can think of the model itself as the solution and those as intermediate steps.
Data (1/1) See above.
Proposed Solution (1/1) You NEED to do more than linear regression and OLS. You should try XGBoost or other regression models and compare them and use cross validation. I will not mark you down for this now, but if you do not use multiple regression models (that aren't just OLS/linear regression) then I will take off points.
Metrics (0.75/1) This is good!
Ethics & Privacy (1/1) This is good.
Team Expectations + Timeline (1/1) Also good.
You can reply to this feedback below. Contact me anytime if you want help improving your project or have any questions at all!