Linear regression doubling time

Click For Summary
SUMMARY

The discussion centers on calculating the doubling time in a linear regression context using the equation y=mx+c, specifically with a logarithmic transformation of the y-axis. The slope (m) is derived as 13/70 from the values 9 and 2.5 over a period of 35 years. The doubling time is determined by the relationship between log y1 and log (2*y1), leading to the formula [(t2)-(t1)] = log2/(13/70). The conversation also clarifies the use of a semilog graph where the y-axis is logarithmic, confirming that the constant c is 10^2.5.

PREREQUISITES
  • Understanding of linear regression equations (y=mx+c)
  • Familiarity with logarithmic functions and transformations
  • Knowledge of semilog and log-log graphing techniques
  • Basic grasp of the concept of doubling time in mathematical contexts
NEXT STEPS
  • Study the derivation of doubling time in exponential growth models
  • Learn about logarithmic transformations in data analysis
  • Explore semilog and log-log graphing techniques in statistical software
  • Investigate the implications of slope in linear regression analysis
USEFUL FOR

Mathematicians, data analysts, and anyone involved in statistical modeling or growth analysis will benefit from this discussion, particularly those interested in linear regression and logarithmic transformations.

yecko
Gold Member
Messages
275
Reaction score
15
Homework Statement
see image
Relevant Equations
y=mx+c
1619617808719.png

nielsen-law-bandwidth-growth-1983-2019.png

y=mx+c
by z=log y axis
m=9-2.5 / 35 = 13/70
z=13/70 * (t-1983) + 2.5
log y = 13/70 * (t-1983) + 2.5 #

doubling time: t1=y1, t2=y2=2y1
log y1 = 13/70 * (t1-1983) + 2.5 ---{1}
log (2*y1) = 13/70 * (t2-1983) + 2.5 ---{2}
{2}-{1}: log2=13/70 * [(t2)-(t1)]
[(t2)-(t1)] = log2/(13/70) #

for log scale, can I log only y axis? for the constant c, is it 2.5 or 10^2.5?
is the doubling time calculated by simultaneous equation like my attempt?

Thank you.
 
Physics news on Phys.org
yecko said:
y=mx+c
by z=log y axis
m=9-2.5 / 35 = 13/70
z=13/70 * (t-1983) + 2.5
log y = 13/70 * (t-1983) + 2.5 #
You are leaving us guessing what you did. A straight line would be ##\ \log_{10} y = m x + c \ ## or ## y = 10^{c} \; \left (10^{x} \right )^m ##

##m=9-2.5 / 35 \ne 13/70 \ \ ## but ## \ \ m= (9-2.5) / 35 = 13/70 ##

And yes, in 70 years ##\log_{10} ## increases by 13 (actually: a little less !), so y becomes a factor 1013 bigger.

Doubling time is when ##\ \ 13/70 * t = \log_{10} 2 \ \ ##. I wouldn't call that solving simultaneous equations...

yecko said:
for log scale, can I log only y axis?
What do you mean ? The verb logging means something else 13
 
  • Like
Likes   Reactions: FactChecker and yecko
yecko said:
for log scale, can I log only y axis?
A graph can be semilog, such as log y vs x, or log-log, such as log y vs log x. The graph you showed is semilog, with t on the horizontal axis, and log(bits per sec) on the vertical axis.

Is that what you're asking?
 
  • Like
Likes   Reactions: FactChecker

Similar threads

  • · Replies 9 ·
Replies
9
Views
1K
  • · Replies 5 ·
Replies
5
Views
3K
Replies
15
Views
3K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 15 ·
Replies
15
Views
6K
  • · Replies 2 ·
Replies
2
Views
5K