Business Administration > QUESTIONS & ANSWERS > University of Southern California BUAD 425425HW2 (All)

University of Southern California BUAD 425425HW2

Document Content and Description Below

1. What percentage of songs in the dataset made it to the Top 10 from the training set? (To the nearest percent) 1060/7201 = 0.147202 or 15% 2. Without building a model, if you had to guess whether... a new song was going to make it to the Top 10, what would you guess? How often would you correctly identify a Top10 Hit in the test set? Is this a good strategy for a record company to follow when deciding whether or not to invest in artists? Justify your answer A: I think this is not a good strategy. I can only identify few new songs in the test setting. And the rate for correctly identify the top 10 song is only about 14%, which means using this method the record company might have 86% loss their money. 3. Build a logistic regression model to predict Top 10 using all the variables EXCEPT: Year; Songtitle; Artistname; SongIDArtistID Why do you think we are excluding these variables? A: Firstly these factors are all category variable that jmp cant predict. Secondly, these variables are irrelevant to prediction of new songs. So we need exclude them. 4. What is the RSquare of your model? (Use 2 decimals) [Show More]

Last updated: 2 years ago

Preview 1 out of 3 pages

Buy Now

Instant download

We Accept: