Skip to content

This is a final project for a Time Series course. My professor told me I could further work on it.

Notifications You must be signed in to change notification settings

e-paj/Time-Series-Project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

59 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Forecasting Kpop Popularity

This is a final project for a Time Series course. My professor told me I could further work on it. Hence, this!!

Facts:

  • I gathered my data from Google Trends. It tracks the popularity of the search term over time. The maximum is 100.
  • For my project, I used the term "K-pop" and the range of my data was from January 2010 to December 2020.
  • Overall, I have 132 observations to use.
  • I used R to analyze my data.
  • (Unfurtunately, I have not saved a copy of my professor's comments about my project, but I do remember some of it.)
  • The final R code I used is Project479Part4.

What I have done so far:

  • The purpose of course is to know what model could best represent my data and use the model to forecast the popularity of the term in the next 50 years.
  • In order to know which model is best, I used six forecasting performance measures to compare each model. These include MSE, RMSE, BIC, etc..
  • I initially tried three ARIMA models of which ARIMA(0,1,3) model has the best performance.
  • I then tried six seasonal ARIMA models of which SARIMA(0,1,3)(0,1,1)12 model has the best performance.
  • Then I included a drift to each model that has the best performance of which ARIMA(0,1,3) with drift model has the best performance.
  • Lastly, from the three models I have I used the measures to pick one.
  • There are 4 R files as I keep making new R code everytime I learn something new in class. As well, as I kept the old code because I want to compare my new code with the old ones and see what are the differences.

What I can do:

  • My professor told me I could compare the popularity with other music genres like Jpop.
  • Also, I was told to be more precise about how I got my model in the end.

What I am currently doing:

  • Currently working on the Jpop data. Trying to see which model would suit it. ✔
  • I am now comparing both data and see how each is doing in the last few years. ✔

Credits:

  • Miller, Ryan & Schwarz, Harrison & Talke, Ismael. (2017). Forecasting Sports Popularity: Application of Time Series Analysis. Academic Journal of Interdisciplinary Studies. 6. 10.1515/ajis-2017-0009.
    • This paper was used as a guide to write my paper, Project_Report.

About

This is a final project for a Time Series course. My professor told me I could further work on it.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages