Probability and Statistical Inference. Robert Bartoszynski
Чтение книги онлайн.

Читать онлайн книгу Probability and Statistical Inference - Robert Bartoszynski страница 22

Название: Probability and Statistical Inference

Автор: Robert Bartoszynski

Издательство: John Wiley & Sons Limited

Жанр: Математика

Серия:

isbn: 9781119243823

isbn:

СКАЧАТЬ alt="images"/> and images (or images and images); this follows again from De Morgan's laws.5

      It is important to realize that closure under countable operations is stronger than closure under any finite operations. This means that there exist classes of sets that are fields but not images‐fields. This is illustrated by the following example:

      Example 1.19

      Let images and let images be the class of all subsets images of images such that either images or images is finite. Then images is a field but not a images‐field. First, if images then images because the definition of images is symmetric with respect to complementation. Next, if images and images are both in images, so is their union. If images and images are both finite, then images is finite and hence belongs to images. On the other hand, if either images or images (or both) are finite, then images is also finite because it is contained in images and also in images.

      Thus, images is a field. However, images is not a images‐field. Let images be the set consisting only of the element images (i.e., images). Clearly, images. Take now images. This is a countable union of sets in images that is not in images since the set of all even numbers is not finite, nor does it have a finite complement.

      Typically, it is easy to determine that a class of sets is a field, while direct verification that it is a images‐field can be difficult. On the other hand, it is sometimes easy to verify that a class of sets is a monotone class.

      Theorem 1.4.2 A images‐field is a monotone class. Conversely, a field that is a monotone class is a images‐field.

      Proof: To prove this theorem, assume first that images is a images‐field, and let images be a monotone sequence of elements of images. If images then images, whereas if images then images So images is a monotone class. On the other hand, let images be a monotone class and a field, and let images be an arbitrary sequence of elements of images. Put images. Then since images is a field, and also images images for every images. Further, since images is a monotone class, images. However, images, so images is a images‐field, as asserted.

      The last in this series of concepts is that of the minimal field (or images‐field, or monotone class) containing a given set or collection of sets. We begin with some examples.

СКАЧАТЬ