Regression
First enter the data set into the data editor: press 2nd STAT, press F2 to select EDIT, then enter the data. Put the input data (independent variable) in the xStat column and the output data (dependent variable) in the yStat column.
After data entry, plot the data to get a good feel for the regression problem at hand. Press 2nd STAT, then press F3 to select PLOT. Press F1 to select PLOT1. Turn PLOT1 ON (highlight On and press ENTER). Press GRAPH, press F3 to zoom, press MORE, and then press F5 to select ZDATA. The displayed example shows the growth of yeast; hence, for our example we conduct a logistic regression.
Press 2nd STAT to access the statistics menu. Press F1 to access the CALC menu. The TI-86 can perform a regression to find the equation of best fit for nine different types of equations: linear (LinR), logarithmic base e (LnR), exponential (ExpR), power (PwrR), sinusoidal (SinR), logistic (LgstR), quadratic (P2Reg), cubic (P3Reg), quartic (P4Reg). We emphasize that only the user can determine the type of function to use.
For our illustration, we find the logistic equation of best fit. Press MORE, then press F3 to select LgstR to conduct a logistic regression. You are now sent to the home screen with LgstR followed by a blinking cursor.
We now inform the TI-86 where the data are located and where we want the regression function saved. Here, the data are in xStat (inputs), yStat (outputs), and we will save the logistic equation in the equation editor as function y1. Accordingly, press
2nd LIST F3 MORE (sufficient times to locate xStat) F2
,
2nd LIST F3 MORE (sufficient times to locate yStat) F3
,
2nd CATLG-VARS F3
locate y, then ENTER
1 ENTER
The logistic information screen is then displayed (see below, left screen). After the regression function is determined, we plot the data with the graph of the regression equation. To view the plot, press GRAPH F5. If desired, press F2 and adjust the window, then press F5 to select GRAPH. For example, the graph at right below uses yMin = -150 so that the graph menu does not obscure the lower part of the graph.
If the plot shows the graph of an unwanted function, go to the equation editor and deselect it by placing the cursor over the highlighted = for that function, then press ENTER.
Note All regressions on the TI-86 are done in the same manner as the example given above. The only difference from one problem to another is the data set used and the equation selected for the regression.