- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
Need some statistics help - regression forecasting
Posted on 12/9/13 at 8:04 pm
Posted on 12/9/13 at 8:04 pm
I've got data for the past few years...applied a trend line and got an Rsquared of .994. Can I use this formula to forecast 2014 and 2015?
I feel like I'm missing something, but I have never claimed to be strong in statistics.
MTB, please help
I feel like I'm missing something, but I have never claimed to be strong in statistics.
MTB, please help
Posted on 12/9/13 at 8:07 pm to lynxcat
quote:
applied a trend line and got an Rsquared of .994
You aren't using live data then. Must be a homework problem.
quote:
Can I use this formula to forecast 2014 and 2015?
You use anything you like to make forecasts. The question is whether they'll be any good after the fact.
Posted on 12/9/13 at 8:11 pm to lynxcat
That's a loaded question. But assuming the prior trend is representative going forward, just pull the equation off of the graph, sub in x for each time period, and it generates y.
If your equation is a linear fit, you can automate it by using the SLOPE() and INTERCEPT() functions in Excel to get the constants without having to read off the graph. If those aren't the exact formula names, it is something like that.
If your equation is a linear fit, you can automate it by using the SLOPE() and INTERCEPT() functions in Excel to get the constants without having to read off the graph. If those aren't the exact formula names, it is something like that.
Posted on 12/9/13 at 9:45 pm to foshizzle
quote:
You aren't using live data then. Must be a homework problem.
This is live data.
quote:
You use anything you like to make forecasts. The question is whether they'll be any good after the fact.
This is my bigger question. The trend line appears quite reflective of what business results I would expect over the near-term.
Posted on 12/9/13 at 9:48 pm to Bayou Tiger
quote:
That's a loaded question. But assuming the prior trend is representative going forward, just pull the equation off of the graph, sub in x for each time period, and it generates y.
This is what I thought -- but I having trouble converting into an Excel formula.
MrExcel is saying to use the POWER function. It is a polynomial.
Posted on 12/9/13 at 10:10 pm to lynxcat
quote:
MrExcel is saying to use the POWER function. It is a polynomial.
Hopefully you didn't use a 12th degree polynomial just to find a good level of regression. You might get some use out of that for interpolation, but for extrapolation look out.
Posted on 12/9/13 at 10:27 pm to Bayou Tiger
quote:
Hopefully you didn't use a 12th degree polynomial just to find a good level of regression. You might get some use out of that for interpolation, but for extrapolation look out.
2nd degree poly
Posted on 12/9/13 at 10:28 pm to lynxcat
I just had another piece of similar data come out with a Rsquared = .992
I thought that only happened in textbooks.
I thought that only happened in textbooks.
Posted on 12/9/13 at 11:22 pm to lynxcat
What is your adjusted Rsquared?
An Rsquared value that high makes me think that multicollinearity is probably an issue in your regression.
Your model might be good for making raw predictions, but be extremely careful about using the coeffecients to explain any of the model.
Edit: what software are you using for your analysis?
An Rsquared value that high makes me think that multicollinearity is probably an issue in your regression.
Your model might be good for making raw predictions, but be extremely careful about using the coeffecients to explain any of the model.
Edit: what software are you using for your analysis?
This post was edited on 12/9/13 at 11:24 pm
Posted on 12/10/13 at 12:37 am to tigercross
quote:
What is your adjusted Rsquared?
Was on my way down to post this.
How many variables are being included in your regression model? The more you use the higher your r-squared but that doesn't necessarily mean that the improvement in R-squared is worth the increasing complexity.
Popular
Back to top
Follow TigerDroppings for LSU Football News