How do you find a linear model when given a table of values?

1 Answer
Jan 5, 2015

If it is 2 sets of values, You could use the Least squares linear regression method:
Select one set of values as x and the other set as y. Then put your calculator into double stat mode, and enter all the values as (x,y) pairs. Then push a for gradient and b for y intercept and the equation is linear of for y = ax + b.
This is only a best fit linear approximation and precisely how close it fits the actual data is reflected by the coefficient of correlation r. he closer r is to 1, the more good is the linear approximation.

If it is only 1 set of values x, then take the log of each value and plot the graph of log x vs x. If the relationship is linear to start with, then the graph will be linear and one may use laws of logs to write it in standard linear form. You can then get the gradient and intercepts from the graph.