Calculus and Statistics

Integrals

1 The Basic Concepts of Function and Probability

1.1 SETS AND FUNCTIONS

Definition 1. A set is any well-defined collection of objects. By well-defined we mean that we can tell what objects are in the collection and what objects are not in the collection. Any member of a set is called an element, or point, of the set.

Example 1. The collection of people who own a home within the city limits of Chicago is a set. Each person owning a home within the city limits of Chicago is an element of this set.

Example 2. If a group of students take an examination, then the collection of scores obtained by the students forms a set. Each individual score is an element of the set.

Example 3. A deck of data cards for use in a computer program forms a set. Each card is an element of the set.

Certain or all, of the elements in one set may be related in some way to certain, or all, of the elements of another set. This point is illustrated in the following examples.

Example 4. Let S be the set of all people. Then the phrase is the parent of relates each element of S, that is, each person, to those elements of S (persons) of which he is the parent. If x is a person who is not the parent of anyone, then "x is the parent of y" will not be satisfied for any person y.

Example 5. Let S be a set of students who took an examination and T be the set of scores obtained by the students. Then the phrase has the score assigns some element of T to each element of S.

Example 6. Let R be the set of real numbers. Then the rule f(x) = 2x³ assigns to each real number x another real number f(x) which is twice the cube of x.

Although the phrase is the parent of (Example 4) relates some people to no one at all, it also relates those who are the parents of several children to more than one person. In Example 6, however, not only does the rule f(x) = 2x³ relate each real number x to some real number, but x is related to a unique real number 2x³. Given x, there is no choice as to what f(x) is. When each element of one set is related to one and only one element of another set (which may also be the same as the first set), then we say that we have a function from the first set into the second set. More formally, we make the following definition:

Definition 2. A rule, phrase, or relationship, which assigns to each element of a set S one and only one element of a set T is said to be a function from S into T.

Thus, the rule in Example 6 is a function, while the phrase in Example 4 does not give a function.

Consider Example 5 again. The phrase has the score is a function from S into T since each student has one and only one score. If a student s has a score t, then we may, if we wish, represent this fact by means of the ordered pair (s, t).* More generally, we may represent the function has the score of Example 5 by all ordered pairs (s, t), where s is a student and s has the score t.

Definition 3. If S and T are any two sets, then any object of the form (s, t), where s is an element of S and t is an element of T, is said to be an ordered pair with s as its first coordinate and t as its second coordinate.

An ordered pair is a pair of elements, one from each of two sets, where the order in which the elements are given is important. An ordered pair is the opposite of an unordered pair, that is, a pair of elements, one from each of two sets, where the order is not taken into account. From the unordered pair containing, say 1 and 2, we can form two ordered pairs: (1, 2) and (2, 1).

If some rule relates each element of a set S to a unique element of a set T, then the rule gives rise to the set of all ordered pairs (s, t) such that s and t are elements of S and T, respectively, and t is related by the rule to s. Moreover, since no element of S is related to more than one element of T, but each element of S is related to some element of T, each element of S will appear as a first coordinate once and only once in the ordered pairs that the rule determines. The collection of ordered pairs determined by such a rule (function) is a subcollection of the set of all ordered pairs that can be formed with an element of S in the first coordinate and an element of T in the second coordinate. This inspires the following definition.

Definition 4. Let S and T be any sets. Then the Cartesian product, or simply the product, of S with T is defined to be the set of all ordered pairs (s, t) such that s and t are elements of S and T, respectively. We denote the product of S with T by S × T.

If f is a function from a set S into a set T, then f determines a particular kind of subcollection of S × T, specifically one in which each element of S appears as a first coordinate once and only once. On the other hand, if we begin with a subcollection of S × T having the property that each element of S appears as a first coordinate exactly once, then this subcollection itself determines a function, namely, the function which relates the element s of S to that element t of T such that (s, t) is the only point of the subcollection which has s as a first coordinate. Functions, therefore, can be considered either as rules relating the elements of one set S to elements of a set T, or as a subcollection of S × T. In sum, we can say:

Characterization of functions in terms of ordered pairs. A function f from a set S into a set T is a collection of elements of S × T such that each element of S appears as a first coordinate in f once and only once.*

We may also indicate that (s, t) is an element of f by writing t = f(s).

Example 7. Suppose S and T both consist of the elements 1, 2, and 3. Braces { } are customarily used to set off the elements of a set. Hence an equivalent form of the first sentence of this example is "Suppose S = T = {1, 2, 3}."

Then S × T contains all the elements in the array shown in Table 1.

Table 1

Each of the following are functions from S into T. Each element of S appears as a first coordinate once and only once in each function. Observe that this is equivalent to saying that each function contains one and only one element from each row of Table 1.

f1 = {(1, 1), (2,2), (3,3)},

f2 = {(1, 3), (2, 3), (3, 1)},

f3 = {(1, 1), (2, 1), (3, 1)}.

These are certainly not all the functions from S into T; in all, there are 27 such functions. Note that f1 can also be characterized by the rule f1(s) = s for each element s of S. The function f3 can be characterized by the rule f3(s) = 1 for each element s of S. We may also specify f2 by stating that

Example 8. Let R be the set of real numbers. Then R × R is the set of all ordered pairs (x, y), where x and y are both real numbers. The reader should recognize R × R as the ordinary coordinate plane. If the subset (= subcollection) f of R × R is a function, then each real number x must appear as a first coordinate in f once and only once. If x0 is a fixed, real number, then the line whose equation is x = x0 is a line parallel to the y-axis of the coordinate plane (Fig. 1). There is exactly one point (x0, f(x0)) of f whose first coordinate is x0; hence there is exactly one point of f on the line x = x0.

On the other hand, if f is a subset of R × R such that each line of the form x = x0 meets f in exactly one point, then f is a function. If some line x = x0 does not meet f, then x0 does not appear as a first coordinate in f; and if some line x = x0 meets f in more than one point, then f contains at least two points with first coordinate x0.

Fig. 1

Example 9. Let S and T be as in Example 7, and

g1 = {(1, 2), (1, 3), (2, 3), (3, 1)},

g2 = {(1, 3), (3, 2)}.

Then g1 and g2 are not functions from S into T. In the case of g1, the element 1 of S appears as a first coordinate twice; hence g1(1) is not clearly defined. On the other hand, 2 does not appear as a first coordinate in g2; hence g2(2) is not defined at all.

The following terminology is standard; we include it for the sake of completeness.

Definition 5. If f is a function from a set S into a set T, we call S the domain of f and T the range of f. The image of f is defined to be the set of all elements t of T such that t = f(s) for some element s of S.

Example 10. In Example 7, the domain and range are both {1, 2, 3} for each of the functions f1, f2, and f3. The image of f1 is {1, 2, 3}, of f2 is {1, 3}, and of f3 is {1}. The image of f is merely the set of elements of the range which appear as second coordinates in f.

EXERCISES

1. We have denoted certain sets by listing their elements between braces, for example f1 = {(1, 1), (2, 2), (3, 3)}. We might also have denoted a set by giving an arbitrary element of the set together with a condition that the element must satisfy to be in the set, all written in the following format: {x | condition that x must satisfy to be in the set}. Thus {y | y is a house} is the set of all houses. Express verbally each of the following sets.

a) {1, 45}

b) {a, b, 6, 7, 9}

c) {{1}}

d) {1, {1}}

e) {x | x is an animal}

f) {w | w is a citizen of Canada}

g) {y | y is an Indian and y lives in Iowa}

h) {z | z is an integer divisible by 2}

2. A set S is said to be a subset of a set T if each element of S is an element of T. Thus {1, 2} is a subset of {1, 2, 3}. A function f from a set S into a set T is a special kind of subset of S × T. We use S ⊂ T to denote that S is a subset of T. Which of the following statements are true? If a statement is true, prove it; if it is false, try to find an instance in which the statement should apply, but does not. In the following, S, T, and W are sets.

a) If S ⊂ T and T ⊂ W, then S ⊂ W.

b) If S ⊂ W, then W ⊂ S.

c) S ⊂ S

d) If S ⊂ T, but T is not a subset of W, then S is not a subset of W.

3. Write out all the elements of the following products. Try to arrange the elements in an array similar to that given in Table 1.

a) {1, 2} × {3, 4}

b) {a, b, c} × {6,7}

c) {q} × {t}

d) {1, 2, 3, 4} × {5, 6, 7, 8}

e) {A, B, C, D} × {A, B, C, D}

4. Find four distinct functions from S into T, where S and T are given as in each of the following. It may help to use the arrays constructed in Exercise 3. Compute the range, domain, and image of each function found.

a) S = {a, b, c} and T = {6, 7}

b) S = {1, 2, 3, 4} and T = {5, 6, 7, 8]

c) S = T = {A, B, C, D}

5. Let S be the set of living human beings. Which of the following phrases define a function from S into S? If a phrase fails to define a function, explain why it fails.

a) Is the cousin of

b) Is the father of

c) Is the same age as

d) Has as mother

If S were the set of all human beings who are living or have ever lived, would any of the phrases (a) through (d) define a function from S into S?

6. Which of the following define functions from the set R of real numbers into R? Indicate those functions whose image is the entire set of real numbers.

a) f(x) = x

b) g(y) = y + 2

c) h(w) = ±3w

d) f(x) = x¹/²

e) g(u) = u²

1.2 THE NOTION OF PROBABILITY

The general purpose of probability theory is to make more mathematical such statements as very likely and not much chance. Given some event E, we wish to assign a number to E, the probability of E, which will measure in some suitable fashion the chance that E will occur. We would also like to know how to manipulate probabilities once they have been assigned. For example, given the probabilities of the events E and E′, we would like to be able to derive the probability of "either E occurs, or E′ occurs."

We formulate the following definitions to help make the discussion more precise.

Definition 6. An experiment is a particular procedure to be performed, or a set of circumstances to be present simultaneously. The particular procedure to be performed, or the set of circumstances to be present, must be clearly defined.

A trial is one particular run of an experiment; that is, a trial is one actual performance of the procedure specified by an experiment, or a particular situation in which all the circumstances called for by the experiment are present.

A simple event is a possible outcome of a particular trial.

A sample space is the set of all simple events associated with an experiment.

The following examples illustrate the concepts presented in Definition 6.

Example 11. Taking a particular coin, tossing it into the air, and letting it come to rest on the floor is an experiment. A particular toss of the coin in accordance with the directions is a trial. The simple events associated with the experiment would be heads and tails; thus the sample space is {heads, tails}.

If the experiment were to flip the coin twice, then heads could occur on both the first and second tosses, or heads could occur on the first toss and tails on the second toss, etc. If H and T represent heads and tails, respectively, then the sample space for this experiment is

S = {HT, TH, HH, TT}.

In this latter experiment the occurrence of exactly one head is equivalent to the occurrence of one of the simple events HT and TH. We can associate the event one occurrence of heads with the subset {HT, TH} of the sample space S.

Definition 7. An event is any subset of a sample space.

Example 12. If the coin of Example 11 is flipped twice, then {HT}, {HH}, {HH, TT}, and S itself are all events. The event {HT, TH, HH} can be thought of as the "nonoccurrence of TT, or as the occurrence of at least one head."

Two events A and B may be mutually exclusive. Informally, this means that A and B cannot occur together. In terms of a sample space, it means that A and B share no simple events in common; that is, their intersection is empty.

Two events may also be independent, that is, the occurrence or nonoccurrence of one of the events in no way affects the occurrence or nonoccurrence of the other event. More will be said about mutually exclusive and independent events later.

We may also "build other events from two given events A and B." In particular, we make the following definition.

Definition 8. Let A and B be two events (subsets of a sample space S). Then A ∩ B, the intersection of A and B, is defined to be

{s | s is an element of S in both A and B}.

A ∩ B is the event A and B. The event A ∪ B, the union of A and B, is defined to be

{s | s is an element of S in either A or B};

A ∪ B is the event either A or B (, the complement of A, is defined to be

{s | s is an element of S, but not an element of A};

is the event not-A.

Fig. 2

Example 13. In Example 12, {HT, TH} ∪ {HH} = {HT, TH, HH}. This may be interpreted to mean, If either one head and one tail occurs, or two heads occur, then two tails do not occur. Note that {HT, TH, HH} is the complement of {TT}. The events {HH, TT} and {HT} have no elements in common; therefore their intersection is the empty set, the set which contains no elements.

Sample spaces furnish us with a precise way to formulate events and enable us to use the machinery of set theory in the development of the theory of probability. Nevertheless, events can still be thought of informally, that is, by merely considering what we wish to consider without explicit reference to a sample space. For example, rain can be thought of as an event without reference to a sample space. Once a suitable sample space has been constructed, however, then rain will be the subset of all elements of the sample space which involve rain.

Example 14. Let A and B be the events of wind and rain, respectively. Then A ∩ B is the event of wind and rain together, while A ∪ B is the event of not-wind. Note that no sample space has been mentioned explicitly.

Definition 9. Two events (subsets of the sample space S) A and B are said to be equivalent if A = B, that is, if A and B contain the same elements of S.

Less formally, we call two events equivalent if each occurs whenever the other occurs.

Example 15. Let A, B, and C be events. Then A ∩ (B ∪ C) is the event of A occurring together with either B or C. In terms of set theory, it is the set of elements (of the sample space) which are both in A and either B or C (Fig. 3). The event (A ∩ B) ∪ (A ∩ C) is the event of A occurring together with B, or A occurring together with C. Set-theoretically, this is the set of elements which are either both in A and B or both in A and C; this again is the set represented in Fig. 3. Thus, the events A ∩ (B ∪ C) and (A ∩ B) ∪ (A ∩ C) are equivalent.

Fig. 3

Given two events A and B, we can define an important event associated with these events which is not, however, a subset of the sample space of A and B.

Definition 10. Given events A and B, the event "A given B," or A if B is denoted by A | B. We may call A | B, the event A conditioned by B.

Example 16. If A and B are rain and wind, respectively, then A | B is rain if there is wind, while B | A is wind if there is rain. Note that in A | B, wind is a given event; we may read A | B as A given B, or A if B.

EXERCISES

1. Let A, B, and C be the events of rain, wind, and lightning, respectively. Write out in words each of the following events.

b) B ∩ C

∩ C

d) A ∪ C

f) A ∩ (B )

g) (A ) ∩ B

h) A

i) B | (A ∩ C)

j) A ∩ B

2. Letting A, B, and C be as in Exercise 1, write symbolically each of the following events.

a) Rain if it is windy and there is lightning

b) Rain without lightning

c) Lightning with neither wind nor rain

d) Lightning if there is neither wind nor rain

e) Lightning and rain without wind

3. In each of the following an experiment is described. Devise a sample space to represent the outcomes of each experiment.

a) A coin is flipped three times.

b) A die is tossed twice.

c) Two dice are tossed together.

d) One slip is drawn from an urn which contains ten slips numbered 1 through 10.

e) Two slips are drawn one at a time from an urn containing ten slips numbered 1 through 10.

4. Let A, B, and C be events (in a sample space S). Prove that the two events in each of the following pairs of events are equivalent. Do this by showing that the sets contain the same elements; a diagram such as that in Fig. 3 may be used for the proof. Write out the meaning of each of the events in words; if done properly, it should be clear from the verbal description of the events that they are equivalent.

a) (A ∩ B) ∩ C; A ∩ (B ∩ C)

b) A

c) (B ∪ C) ∩ A; (B ∩ A) ∪ (C ∩ A)

5. Formulate three distinct experiments that might be performed with an ordinary deck of cards. For each experiment, formulate a sample space.

6. We have not yet defined what is meant by the probability of an event; however, the reader should already have some intuitive ideas about the properties of probabilities. If A is any event, we let P(A) denote the probability of A. In each of the following, A and B are events. Decide which of the following statements are true and which are false. Justify each of your assertions.

a) P(A ∪ B) is at least as large as P(A).

b) P(A ∩ B) is no larger than P(B).

c) P(A ∪ B) = P(A) + P(B).

d) P(A ∩ B) is less than or equal to P(A)P(B).

e) If A and B are mutually exclusive, then P(A ∩ B) will have the smallest value permitted for a probability.

f) If A is twice as probable as B, then P(A) = 2P(B).

1.3 THE BASIC LAWS OF PROBABILITY

We would like to define the probability P(A) of an event A in such a manner that if the experiment associated with A is performed n times, then A should occur about nP(A, and the coin is flipped 1000 times, then heads should occur about 500 times. It follows then that if we wish to find the probability of an event A, we could perform the appropriate experiment n times. If A occurs m times out of the n trials, then we may estimate P(A) to be m/n.

Example 17. A balanced coin is flipped in an unbiased manner 1000 times and 495 heads are obtained. Then we can estimate P(heads) to be .495.

Example 18. The weatherman finds that 80 out of 100 days on which the temperature in Chicago was above 75° saw rain. The probability of rain in Chicago on a day when the temperature is above 75° can therefore be estimated as .8.

Definition 11. If some particular experiment is performed n times and the event A occurs on m of the n trials, then m/n is called the relative frequency of A. We shall denote the relative frequency of A by R(A).

We have said that R(A) could be used to estimate P(A). However, this relative frequency approach to probability has certain limitations.

First, in order to obtain a reasonably good estimate of a probability, a fairly large number of trials would have to be performed. As an extreme example, if we flip a coin only once, then we obtain either heads or tails. Using just one flip, we would then estimate the probability of heads to be either 0 or 1, depending on whether heads or tails had come up. About the only thing that one flip might tell us is that whatever was flipped has a nonzero probability.

, it is not impossible to flip 1000 consecutive tails. Thus, in computing an estimate of the probability using the relative frequency approach, we can never be really sure that the unusual has not occurred and the estimate is far off the mark. What does increase with the number of trials is the probability that our estimate will approach the true probability.

In certain instances, it might be hard to get enough information to make an intelligent estimate of the probability. The following example illustrates this point.

Example 19. A veterinarian wishes to know the probability that a certain species of fox will develop rabies. The veterinarian may, however, be able to study only a very limited number of foxes. Moreover, the fact that the foxes are observed, particularly if the observation removes them from their natural habitat, may grossly distort the veterinarian’s conclusions. For example, all sick foxes may hide in their burrows and hence cannot be observed, or those captured may not contract the disease because they have been removed from its source.

Despite its limitations, we shall accept the relative frequency approach to probability as having some validity. We shall use it to motivate the axioms for a theory of probability.

Fig. 4

Let us assume now that A and B are two events associated with some experiment. Then the outcome of any particular trial of the experiment must fall into one and only one of the following categories: A ∩ B∩ B, A (Fig. 4). That is, A occurs with or without B, or B occurs without A, or neither A nor B occurs. Suppose n trials are performed. We summarize the results as shown in Table 2. We shall use this table to derive important relationships between relative frequencies.

Table 2

We first compute R(A). The event A is equivalent to (A ∩ B) ∪ (A ). Since A ∩ B occurs p times and A occurs r times, A occurs p + r times. This means that

Similarly,

We therefore have R(A) + R) = (p + q + r + s)/n = n/n = 1. Hence we have proved

Since p, q, r, s, and n are all nonnegative integers, we may also say

Now consider A ∪ B. Since A ∪ B is equivalent to

(A ∩ B) ∪ (A ∩ B),

A ∪ B occurs p + q + r times. We therefore have

As we have seen, R(A) is (p + r)/n, while R(B) is (p + q)/n. Thus, we can say

But since p/n = R(A ∩ B), using (5), we find that (6) becomes

or

Finally, consider R(A ∩ B). We have R(A ∩ B) = p/n. Dividing R(A ∩ B) by R(A), we obtain

Now p + r is the number of trials in which A occurs, and p is the number of trials in which both A and B occur. Therefore p/(p + r) is the relative frequency of B in those trials in which A occurs; that is, p/(p + r) is nothing other than R(B | A). Equation (9) thus becomes

which is equivalent to

If we assume that for a large number of trials, R(A) is a good estimate of P(A), then the probabilities should satisfy the same relationships that relative frequencies satisfy. In particular, motivated by the relationships among relative frequencies derived in this section, we make the following assumptions about probabilities.

Assumption 1. If A is any event, then P(A), the probability of A, is a real number which lies between 0 and 1, inclusive. P(A) = 0 if A is the empty set (an impossible event) and P(A) = 1 if A is the entire sample space (a certain event).

Assumption 2. If A and B are events, then

P(A ∪ B) = P(A) + P(B) – P(A ∩ B).

Assumption 3. For any events A and B,

P(A ∩ B) = P(A)P(B | A) = P(B)P(A | B)

if P(A) and P(B) are not 0. If either P(A) or P(B) is 0, then P(A ∩ B) = 0.

In the next section