4.

1 - Exponential Functions and Their

Graphs
Exponential Functions
So far, we have been dealing with algebraic functions. Algebraic functions are
functions which can be expressed using arithmetic operations and whose
values are either rational or a root of a rational number. Now, we will be
dealing with transcendental functions. Transcendental functions return values
which may not be expressible as rational numbers or roots of rational
numbers.

Algebraic equations can be solved most of the time by hand. Transcendental

functions can often be solved by hand with a calculator necessary if you want
a decimal approximation. However when transcendental and algebraic
functions are mixed in an equation, graphical or numerical techniques are
sometimes the only way to find the solution.

The simplest exponential function is: f(x) = ax, a>0, a≠1

The reasons for the restrictions are simple. If a≤0, then when you raise it to a
rational power, you may not get a real number. Example: If a=-2, then (-2)0.5 =
sqrt(-2) which isn't real. If a=1, then no matter what x is, the value of f(x) is 1.
That is a pretty boring function, and it is certainly not one-to-one.

Recall that one-to-one functions had several properties that make them
desirable. They have inverses that are also functions. They can be applied to
both sides of an equation.

Graphs of Exponential Functions

 The graph of y=2x is shown to the right. Here are
some properties of the exponential function when the base is greater than 1.

 The graph passes through the point (0,1)

 The domain is all real numbers
 The range is y>0.
 The graph is increasing
 The graph is asymptotic to the x-axis as x approaches negative infinity
 The graph increases without bound as x approaches positive infinity
 The graph is continuous
 The graph is smooth

What would the translation be if you replaced every x with -x? It would be a
reflection about the y-axis. We also know that when we raise a base to a
negative power, the one result is that the reciprocal of the number is taken.
So, if we were to graph y=2-x, the graph would be a reflection about the y-axis
of y=2x and the function would be equivalent to y=(1/2)x.
 The graph of y=2-x is shown to the right. Properties
of exponential function and its graph when the base is between 0 and 1 are
given.

 The graph passes through the point (0,1)

 The domain is all real numbers
 The range is y>0.
 The graph is decreasing
 The graph is asymptotic to the x-axis as x approaches positive infinity
 The graph increases without bound as x approaches negative infinity
 The graph is continuous
 The graph is smooth

Notice the only differences regard whether the function is increasing or

decreasing, and the behavior at the left hand and right hand ends.

Translations of Exponential Graphs

You can apply what you know about translations (from section 1.5) to help you
sketch the graph of the exponential functions.

A horizontal translation may affect the increasing / decreasing (if multiplied by

a negative), the left hand / right hand behavior of the graph, and the y-
intercept, but it won't change the location of the horizontal asymptote.

A vertical translation may affect the increasing / decreasing (if multiplied by a

negative), the y-intercept, and the location of the horizontal asymptote. It will
not change whether the graph goes without bound or is asymptotic (although
it may change where it is asymptotic) to the left or right.
By knowing the features of the basic graphs, you can apply those translations
to easily sketch the new function.

You should now add the exponential graph from the front cover of the text to
the list of those you know.

The natural base e

As x increases without bound, the quantity (1+1/x)^x will approach the
transcendental number e.

The limit notation shown is from calculus. The limit notation is a way of asking
what happens to the expression as x approaches the value shown. The limit is
the dividing line between calculus and algebra. Calculus is algebra with the
concept of limit. People always have this dread of calculus that I can't
understand. The calculus itself is easy. The reason people don't do well in
calculus is not because of the calculus, but because of they are poor at
algebra.

The value for e is approximately 2.718281828. Here is a slightly more

accurate, but no more useful, approximation.

2.71828 18284 59045 23536 02874 71352 66249 77572 47093 69995 95749
66967 62772 40766 30353 54759 45716 82178 52516 64274 27466 39193
20030 59921 81741 35966 29043 57290 03342 95260 59563 07381 32328
62794 34907 63233 82988 07531 95251 01901 15738 34187 93070 21540
89149 93488 41675 09244 76146 06680 82264 80016 84774 11853 74234
54424 37107 53907 77449 92069 55170 27618 38606 26133

When the base is e used, the exponential function becomes f(x) = ex. There is
a key on your calculator labeled e^x. On the TI-8x calculators, it is on the left
side as a [2nd] [Ln]. The exponential function with base e is sometimes
abbreviated as exp(). One common place this abbreviation appears is when
writing computer programs. I mention this so when I write exp(x), you know
what I'm talking about.

Compounded Interest
The amount in your savings account can be figured with exponential
functions. Each period (I'll assume monthly), you get 1/12 of the annual
interest rate (r) applied to your account. The new amount in the account is
100% of what you started with plus r%/12 of what you started with that. That
means that you now have (100%+r%/12) of what you started with. The next
month, you will have the same thing, except it will be based on what you had
at the end of the first month.

Confusing, I know. There is an explanation on page 304 of the text, but the
resulting formula for compound interest is A = P (1+i)n.

A is the Amount in the account. P is the principal you started with. i is the
periodic rate, which is the annual percent (written as a decimal) r, divided by
the number of periods per year, m. n is the number of compounding periods,
which is equal to the number of periods per year, m, times the time in years, t.
The formula I have shown above differs slightly from the formula in the book,
but agrees with the formula that you'll use if you go on to Finite Mathematics
(Math 160). In Finite Mathematics, there is an entire chapter on finance and
the formulas involved.

Continuous Compounding and Growth / Decay

In the old days, there used to be continuous compounding of interest. You
don't find it any more because it is gives the highest return on the investment,
and banks are in business to make money, just like any other for profit
institution.

The model for continuous compounding is A = P ert.

A is the Amount, P is the Principal, r is the annual percentage rate (written as

a decimal), and t is the time in years. e is the base for natural logarithms.

However, the continuous model does make sense for population growth and
radioactive decay. The radioactivity of an isotope doesn't change once a
month at the end of the month, it is continually changing.

The exponential model is y = A ekt,

where y is the amount present at time t. A is the initial amount present, and k
is the rate of growth (if positive) or the rate of decay (if negative).
4.2 - Logarithmic Functions and Their
Graphs
Inverse of Exponential Functions

We stated in the section on exponential

functions, that exponential functions were one-to-one. One-to-one functions
had the special property that they have inverses that are also functions. And,
as many of you said in class, and I'm so glad you remember, one-to-one
functions can be applied to both sides of an equation. They also pass a
horizontal line test.

This section is about the inverse of the exponential function. The inverse of an
exponential function is a logarithmic function. Remember that the inverse of a
function is obtained by switching the x and y coordinates. This reflects the
graph about the line y=x. As you can tell from the graph to the right, the
logarithmic curve is a reflection of the exponential curve.

The table below demonstrates how the x and y values of the points on the
expontential curve can be switched to find the coordinates of the points on the
logarithmic curve.

Point on Corresponding point

exponential curve on logarithmic curve

(-3, 1/8) (1/8, -3)

 (-2, 1/4) (1/4, -2)

(-1, 1/2) (1/2, -1)

(0, 1) (1, 0)

(1, 2) (2, 1)

(2, 4) (4, 2)

(3, 8) (8, 3)

Comparison of Exponential and Logarithmic Functions

Let's look at some of the properties of the two functions.

The standard form for a logarithmic function is: y = loga x

Note, if the "a" in the expression above is not a subscript (lower than the
"log"), then you need to update your web browser.

Exponential Logarithmic

Function y=ax, a>0, a≠1 y=loga x, a>0, a≠1

Domain all reals x>0

Range y>0 all reals

intercept y=1 x=1

increasing when a > 1 when a > 1

decreasing when 0 < a < 1 when 0 < a < 1

asymptote y-axis x-axis

 continuous yes yes

smooth yes yes

Working Definition of Logarithm

In the exponential function, the x was the exponent. The purpose of the
inverse of a function is to tell you what x value was used when you already
know the y value. So, the purpose of the logarithm is to tell you the exponent.

Thus, our simple definition of a logarithm is that it is an exponent.

Another way of looking at the expression "loga x" is "to what power (exponent)
must a be raised to get x?"

Equivalent Forms

The logarithmic form of the equation y=loga x is equivalent to the exponential

form x=ay.

To rewrite one form in the other, keep the base the same, and switch sides
with the other two values.

Properties of Logarithms
loga 1 = 0 because a0 = 1

No matter what the base is, as long as it is legal, the log of 1 is always
0. That's because logarithmic curves always pass through (1,0)

loga a = 1 because a1 = a

Any value raised to the first power is that same value.

loga ax = x

The log base a of x and a to the x power are inverse functions.

Whenever inverse functions are applied to each other, they inverse out,
and you're left with the argument, in this case, x.

loga x = loga y implies that x = y

 If two logs with the same base are equal, then the arguments must be
equal.

loga x = logb x implies that a = b

If two logarithms with the same argument are equal, then the bases
must be equal.

Common Logs and Natural Logs

There are two logarithm buttons on your calculator. One is marked "log" and
the other is marked "ln". Neither one of these has the base written in. The
base can be determined, however, by looking at the inverse function, which is
written above the key and accessed by the 2nd key.

Common Logarithm (base 10)

When you see "log" written, with no base, assume the base is 10.
That is: log x = log10 x.

Some of the applications that use common logarithms are in pH (to measure
acidity), decibels (sound intensity), the Richter scale (earthquakes).

An interesting (possibly) side note about pH. "Chapter 50: Sewers" of the
Village of Forsyth Code requires forbids the discharge of waste with a pH of
less than 5.5 or higher than 10.5 (section 50.07).

Common logs also serve another purpose. Every increase of 1 in a common

logarithm is the result of 10 times the argument. That is, an earthquake of 6.3
has 10 times the magnitude of a 5.3 earthquake. The decibel level of loud
rock music or a chainsaw (115 decibels = 11.5 bels) is 10 times louder
than chickens inside a building (105 decibels = 10.5 bels)

Natural Logarithms (base e)

Remember that number e, that we had from the previous section? You know,
the one that was approximately 2.718281828 (but doesn't repeat or
terminate). That is the base for the natural logarithm.

When you see "ln" written, the base is e.

That is: ln x = loge x
Exponential growth and decay models are one application that use natural
logarithms. This includes continuous compounding, radioactive decay (half-
life), population growth. Typically applications where a process is continually
happening. Now, these applications were first mentioned in the exponential
section, but you will be able to solve for the other variables involved (after
section 4) using logarithms.

In higher level mathematics, the natural logarithm is the logarithm of choice.

There are several special properties of the natural logarithm function, and it's
inverse function, that make life much easier in calculus.

Since "ln x" and "ex" are inverse functions of each other, any time an "ln" and
"e" appear right next to each other, with absolutely nothing in between them
(that is, when they are composed with each other), then they inverse out, and
you're left with the argument.

4.3 - Properties of Logarithms

Change of Base Formula

One dilemma is that your calculator only has logarithms for two bases on it.
Base 10 (log) and base e (ln). What is to happen if you want to know the
logarithm for some other base? Are you out of luck?

No. There is a change of base formula for converting between different bases.
To find the log base a, where a is presumably some number other than 10
or e, otherwise you would just use the calculator,

Take the log of the argument divided by the log of the base.

loga x = ( logb x ) / ( logb a )

There is no need that either base 10 or base e be used, but since those are
the two you have on your calculator, those are probably the two that you're
going to use the most. I prefer the natural log (ln is only 2 letters while log is 3,
plus there's the extra benefit that I know about from calculus). The base that
you use doesn't matter, only that you use the same base for both the
numerator and the denominator.

loga x = ( log x ) / ( log a ) = ( ln x ) / ( ln a )

Example: log3 7 = ( ln 7 ) / ( ln 3 )

Logarithms are Exponents

Remember that logarithms are exponents, so the properties of exponents are

the properties of logarithms.

Multiplication

What is the rule when you multiply two values with the same base together
(x2 * x3)? The rule is that you keep the base and add the exponents. Well,
remember that logarithms are exponents, and when you multiply, you're going
to add the logarithms.

The log of a product is the sum of the logs.

loga xy = loga x + loga y

Division

The rule when you divide two values with the same base is to subtract the
exponents. Therefore, the rule for division is to subtract the logarithms.

The log of a quotient is the difference of the logs.

loga (x/y) = loga x - loga y

Raising to a Power

When you raise a quantity to a power, the rule is that you multiply the
exponents together. In this case, one of the exponents will be the log, and the
other exponent will be the power you're raising the quantity to.

The exponent on the argument is the coefficient of the log.

loga xr = r * loga x

Melodic Mathematics

Some of the statements above are very melodious. That is, they sound good.
It may help you to memorize the melodic mathematics, rather than the
formula.
  The log of a product is the sum of the logs
 The sum of the logs is the log of the products
 The log of a quotient is the difference of the logs
 The difference of the logs is the log of the quotient
 The exponent on the argument is the coefficient of the log
 The coefficient of the log is the exponent on the argument

Okay, so the last two aren't so melodic.

Common Mistakes

I almost hesitate to put this section in here. It seems when I try to point out a
mistake that people are going to make, that more people make it.

 The log of a sum is NOT the sum of the logs. The sum of the logs is the
log of the product. The log of a sum cannot be simplified.
loga (x + y) ≠ loga x + loga y
 The log of a difference is NOT the difference of the logs. The difference
of the logs is the log of the quotient. The log of a difference cannot be
simplified.
loga (x - y) ≠ loga x - loga y
 An exponent on the log is NOT the coefficient of the log. Only when the
argument is raised to a power can the exponent be turned into the
coefficient. When the entire logarithm is raised to a power, then it can
not be simplified.
(loga x)r ≠ r * loga x
 The log of a quotient is not the quotient of the logs. The quotient of the
logs is from the change of base formula. The log of a quotient is the
difference of the logs.
loga (x / y) ≠ ( loga x ) / ( loga y )

4.4 - Solving Exponential and Logarithm

Equations
Inverses

Remember that exponential and logarithmic functions are one-to-one

functions. That means that they have inverses. Also recall that when inverses
are composed with each other, they inverse out and only the argument is
returned. We're going to use that to our benefit to help solve logarithmic and
exponential equations.

Please recall the following facts:

 loga ax = x
 log 10x = x
 ln ex = x
 a loga x = x
 10log x = x
 eln x = x

Solving Exponential Equations Algebraically

1. Isolate the exponential expression on one side.

2. Take the logarithm of both sides. The base for the logarithm should be
the same as the base in the exponential expression. Alternatively, if you
are only interested in a decimal approximation, you may take the natural
log or common log of both sides (in effect using the change of base
formula)
3. Solve for the variable.
4. Check your answer. It may be possible to get answers which don't
check. Usually, the answer will involve complex numbers when this
happens, because the domain of an exponential function is all reals.

Solving Logarithmic Equations Algebraically

1. Use properties of logarithms to combine the sum, difference, and/or

constant multiples of logarithms into a single logarithm.
2. Apply an exponential function to both sides. The base used in the
exponential function should be the same as the base in the logarithmic
function. Another way of performing this task is to write the logarithmic
equation in exponential form.
3. Solve for the variable.
4. Check your answer. It may be possible to introduce extraneous
solutions. Make sure that when you plug your answer back into the
arguments of the logarithms in the original equation, that the arguments
are all positive. Remember, you can only take the log of a positive
number.
Solving Equations Graphically

Sometimes, it is impossible to solve an equation involving logarithms or

exponential functions algebraically. This is especially true when the equation
involves transcendental (logs and/or exponentials) and algebraic components.
In cases like these, it may be necessary to use the graphing calculator to help
find the solution to the equation.

1. Rewrite the equation so that all the terms are on one side
2. Graph the expression
3. Use the Root or Zero function under the Calc menu.

4.5 - Exponential and Logarithmic

Models
Exponential Growth

Function

y = C ekt, k > 0

Features

 Asymptotic to y = 0 to left
 Passes through (0,C)
 C is the initial value
 Increases without bound to right
Notes

Some of the things that exponential growth is used to model include

population growth, bacterial growth, and compound interest.

If you are lucky enough to be given the initial value, that is the value when x =
0, then you already know the value of the constant C. The only thing
necessary to complete the model is to have one additional point on the graph.
Plug in the values for x, y, and C, and solve for k.

Alternatively, almost like cheating, you can put the x-values into List 1, the y-
values into List 2, and choose the ExpReg option on the TI-82 calculator.

Exponential Decay (decreasing form)

Function

y = C e-kt, k > 0

Features

 Asymptotic to y = 0 to right
 Passes through (0,C)
 C is the initial value
 Decreasing, but bounded below by y=0

Notes

Exponential decay and be used to model radioactive decay and depreciation.

Exponential decay models decrease very rapidly, and then level off to become
asymptotic towards the x-axis.

Like the exponential growth model, if you know the initial value then the rest of
the model is fairly easy to complete.

Exponential Decay (increasing form)

Function

y = C ( 1 - e-kt ), k > 0

Features

 Asymptotic to y = C to right
 Passes through (0,0)
 C is the upper limit
 Increasing, but bounded above by y=C

Notes

Exponential decay models of this form can model sales or learning curves
where there is an upper limit. This is done by subtracting the exponential
expression from one and multiplying by the upper limit.

Exponential decay models of this form will increase very rapidly at first, and
then level off to become asymptotic to the upper limit.
Like the other exponential models, if you know upper limit, then the rest of the
model is fairly easy to complete.

The calculator will not fit the increasing model involving exponential decay
directly.

Gaussian Model

Function

y = a exp (-(x-c) / b)2

exp(x) is another way of writing ex

Features

 Asymptotic to x axis to the left and right

 Passes through (0,a)
 a controls the height of the curve
 Centered about x=c
 b controls the spread
 bell shaped

Notes

The Gaussian model is used quite a bit in Statistics to model the distribution of
scores.
The Gaussian model is named after the German mathematician Carl Friedrich
Gauss. This is the same Gauss who developed the Fundamental Theorem of
Algebra. It is symmetric about its mean, x = c. To the right of the origin, it
decreases slowly at first, then more rapidly, and then levels off and become
asymptotic to the x-axis.

Like the Exponential Decay model, the Gaussian model can be turned into an
increasing function by subtracting the exponential expression from one and
then multiplying by the upper limit.

Logisitics Growth Model

Function

y = a / (1 + b e-kx ), k > 0

Features

 Asymptotic to y = a to right,
 Asymptotic to y = 0 to left,
 Passes through (0, a/(1+b) )
 Slow growth, followed by moderate growth, followed by slow growth

Notes

My personal favorite, the logistics model begins with a slow growth, followed
by a period of moderate growth, and then back to a period of slow growth. It
has an upper limit that cannot be exceeded.
The Logistics model can be used to approximate sales and advertising (a little
advertising generates a little growth in sales, more advertising generates
moderate growth in sales, and finally there reaches a point of saturation
where additional advertising benefits little in terms of sales) or population
growth where there is not the capacity for unlimited growth (use the
exponential growth model if you need that).

Logarithmic Model

Function

y = a + b ln x

Features

 Increases without bound to right

 Passes through (1,a),
 Very rapid growth, followed by slower growth,
 Common log will grow slower than natural log
 b controls the rate of growth

The logarithmic model has a period of rapid increase, followed by a period

where the growth slows, but the growth continues to increase without bound.
This makes the model inappropriate where there needs to be an upper bound.
The main difference between this model and the exponential growth model is
that the exponential growth model begins slowly and then increases very
rapidly as time increases.
Several physical applications have logarithmic models. The magnitude of
earthquakes, the intensity of sound, the acidity of a solution.

Earthquakes

R = log I

The Richter scale is used to measure the intensity of an earthquake. The

actual model is a little more complex, but it simplifies to the equation
shown. R is the magnitude on the Richter scale of the earthquake. I is the
intensity of the earthquake measured relative to a reference value. That
reference value is the smallest seismic activity that can be measured, and has
the value I0 = 1.

Every increase of 1 in the Richter scale means the magnitude of the

earthquake is 10 times greater.

Sound Intensity

β = 10 log ( I / I0 )

The sound level, measured in decibels, is given by the formula shown. The
reference value, I0 is the smallest sound intensity which can be heard by the
human ear and is roughly equal to 1x10-16 watts per square centimeter.

There are 10 decibels to a bel. While bel is the actual unit, like meter, liter, or
gram, we use decibel for all practical purposes. An increase of 10 decibels is
equivalent to a sound that is 10 times as strong in intensity. An increase of 20
decibels is equivalent to a sound intensity that is 100 times greater.

Acidity

pH = - log [ H+ ]

pH is a measure of the acidity of a substance. It ranges from 0 to 14, with

acids ranging from 0 to 7, 7 being neutral, and bases ranging from 7 to 14.
The [H+] is the concentration of hydrogen ions, and is measured in moles per
liter. The more hydrogen ions, then smaller the pH (notice the negative sign in
front of the log) and the more acidity the solution is.
Upper Bounds

Notice that the Exponential Growth and Logarithmic models increase without
bound to the right. The Gaussian and Exponential Decay models both
approach the x-axis to the right. Only the Logistics Growth model gives you an
upper bound to the right. The Exponential Decay and Gaussian models can
be made to have an upper bound by subtracting the exponential expression
(thus making it decreasing instead of increasing) from one and multiplying by
the upper bound

4.6 - Exploring Data: Nonlinear Models

This section is a continuation of section 2.6, Exploring Data: Linear Models
and Scatter Plots. Please see that section for the introductory information on
how to enter data into your calculator, create scatter plots, and perform
regression.

TI-82 / TI-83 Models

Model Calculator Equation
Command

Linear LinReg(ax+b) L1, L2 y = ax + b

Quadratic,
At least 3 points QuadReg L1, L2 y = ax2 + bx + c
required

Cubic,
At least 4 points CubicReg L1, L2 y = ax3 + bx2 + cx + d
required

Quartic,
y = ax4 + bx3 + cx2 + dx +
At least 5 points QuartReg L1, L2
e
required

Logarithmic LnReg L1, L2 y = a + b ln x

 Exponential ExpReg L1, L2 y = a bx

Power PwrReg L1, L2 y = a xb

Logistic (TI83 only) Logistic L1, L2 y = c / ( 1 + a e-bx )

Sinusoidal (TI83 only),

At least 4 points SinReg L1, L2 y = a sin (bx + c) + d
required

With the TI-83, you may store the regression equation directly into a y-variable
by specifying the y-variable after the lists where your data is stored. Example:
"LnReg L1, L2, y1" would compute the logarithmic model and save the result
into y1.

TI-85 Models
Model Calculator Equation
Command

Linear LINR y = a + bx

Quadratic,
At least 3
P2REG
points
required
You are given a list of coefficients. These
Cubic,
coefficients are in standard form, so the first
At least 4
P3REG coefficient will be the leading cofficent, and
points
the last coefficient in the list will be the
required
constant.
Quartic,
At least 5
P4REG
points
required
 Logarithmic LNR y = a + b ln x

Exponential EXPR y = a bx

Power PWRR y = a xb

Notice that things are a little different for the TI-85 when it comes to
polynomial regression. The results are not given in equation form with
individual coefficients, but are given as a list of coefficients and stored in the
variable PRegC. If PRegC = { 3, 2, -1, 5 }, then the regression equation would
be y = 3x3 + 2x2 - x + 5. DRREG will correctly draw the regression equation
even if polynomial regression is used.

Least Squares Regression Model

The concept behind fitting a model to data is to find the model that has the
smallest deviation from the actual data. This deviation is the predicted
(estimated by the model) y-value minus the actual y-value. These deviations
are all squared to make sure they don't cancel each other out, and then
added. We're then looking for the smallest sum of the squared deviations, that
is the "Least Squares" model.

TI-82 / TI-83

Note problems 35 and 36 will be much easier if you follow the instructions
here than the instructions in the book.

Assume that you have entered the x-values into list 1 and the y-values into list
2. You have computed whichever model you're working with and have saved
the resulting equation into y1.

The predicted y-values can be found by evaluating y1 at each x-value (list

1): y1(L1)

The sum of the squared deviations can be found using the following
command:
sum ( ( L2 - y1(L1) )2 )
  The "sum" command can be found under the LIST key. After hitting
2nd Stat to get the List menu, arrow right to Math, and choose sum.
 The L1 and L2 commands can be obtained by hitting 2nd 1 and 2nd 2.
 The y-variables like y1 can be found on the TI-82 under the Y-VARS key
(2nd VARS), and on the TI-83 under the VARS key, and then arrow right
to the Y-Vars option.

The model with the smallest sum of the squared deviations is the best model.

TI-85

With the TI-85, this is a much harder task to accomplish. Calculate the
regression equation as normal, then enter the following expression to find the
sum of the squares (extra spaces inserted for readability).
sum seq( ( yStat(x) - fcsty xStat(x) )2, x, 1, dimL xStat, 1)

 sum, seq, and dimL can be found under the LIST key, and the OPS sub
menu.
 The xStat and yStat can be found under the LIST key, and the NAMES
sub menu. They can also be found under the STAT key and the VARS
sub menu, but it is probably easier from list since that's where the rest of
the commands are.
 The fcsty can be found under the CATALOG by pressing the Alpha and
then F keys.

If someone can find an easier way on the TI-85, then please let me know. You
may want to write a short program with that command in it.

Correlation Coefficient

We mentioned the correlation coefficient, r, back in section 2.6. Back then, it

was only for linear models. We can also compute (well, we'll let the calculator
compute) the value for r and r2 for nonlinear models also.

Back in section 2.6, we said that the closer the value of r was to 1 or -1, the
better the model fit the data. For nonlinear models, we often ignore the value
of r and look at the value of r2. The value of r2 can be found by simply squaring
r.

There is always variation in your data. That is, never (almost) is all of the data
exactly the same. The interpretation of r2, from a statistical point of view is the
percent of the variation that can be explained by the model. The higher the
value of r-squared, the better the model explains the data.

The TI-85 does not calculate the value of r2 for polynomial (quadratic, cubic, or
quartic models). For the other models, you may have to square the correlation
(corr variable). The variables necessary to easily compute the value the
correlation coefficient aren't computed when doing polynomial regression, so
finding r or r2 for polynomial regression are going to be extremely difficult for
the TI-85 users. I suggest you find a TI-82 or TI-83 and borrow it long enough
to do that particular regression model.

The model with the highest value for r2 is the best model.