# Have exponential graph, find equation?

1. Dec 11, 2007

### ww2steel

Hi, It's been so long since I took any math classes! Please bear with me. I have sets of three points on various exponential graphs such as 0=100, 30=138, and 60=392 (x and y). I have found manually that this equation is somewhere very close to y=.0025x^2.85+100. I CAN do these manually (plugging in numbers into Excel and modifying them to hit the desired points on a graph), but I have a few HUNDRED of these sets. Is there a faster way? Hopefully that I can put in the 0, 30, and 60 values and Excel can determine the equation? (The built in Excel equations I have used from the graphs so far are completely wrong for what I am doing.)

Please shoot me an email if you can help me! It will help me so much! (Please keep it simple!)

Thanks a lot!
Mike

2. Dec 11, 2007

### ww2steel

The graph

In case it helps to figure out what my non-technical post is talking about. I need to put a line through all three pink dots. The pink dots move all around, but they are all reachable with variations of the above equation. Please help me find an easier way to do this. (Upon reevaluation I have over a thousand of these to do!) Thanks!

#### Attached Files:

• ###### 45mm APCR.jpg
File size:
38 KB
Views:
86
3. Dec 11, 2007

### ww2steel

Infinity at 90°!

Ah, crap, I just realized I have a fourth point - infinity at 90 (degrees). Do I need to throw a cosine in there or something, now I'm fairly lost. My current equations do not go to infinity at any set point, (limit?), nor do I know how to make them do so.

#### Attached Files:

• ###### 45mm APCR.jpg
File size:
34.2 KB
Views:
85
Last edited: Dec 11, 2007
4. Dec 11, 2007

### Office_Shredder

Staff Emeritus
infinity at ninety degrees means your function is probably some variant of secant(x). Just try a*secant(x) + b, or maybe a*secant(x)2+b*secant(x)+c for a,b,c unknown constants. That way you're guaranteed to be able to solve for them algebraically, but the picture may not be what you're looking for

Last edited: Dec 11, 2007
5. Dec 12, 2007

### EnumaElish

I Excel you should be able to fit a power series trendline (on the graph). Right-click on any point of the series, then click "Add trendline." Then choose type.

An alternative is to run a least squares regression using Excel's Data Analysis Toolpak Add-In. (You may have to install Data Analysis TPAI first.)

The easiest linear equation to estimate would be Log[y-100] = a + b Log[x].

y = 100 + c*x^b ===> Log[y-100] = Log[c] + b Log[x] = a + b Log[x], where a = Log[c].