Summary
Highlights
The video introduces the Cumulative Distribution Function (CDF), denoted by capital F(x), for discrete random variables. It distinguishes CDF from the Probability Mass Function (PMF), small f(x), emphasizing that PMF gives the probability of a specific value, while CDF gives the probability of a value less than or equal to a given 'x'. The term 'cumulative' signifies the process of successive additions of probabilities.
To be a valid distribution function, F(x) must satisfy certain conditions: F(x) approaches 0 as x approaches negative infinity, and F(x) approaches 1 as x approaches positive infinity. Additionally, if 'a' is less than 'b', then F(a) must be less than or equal to F(b), indicating that the CDF is a non-decreasing function.
The video provides an example using the number of heads obtained in four coin tosses. It demonstrates how to construct a CDF table by cumulatively adding the probabilities from the PMF. For instance, F(1) is the sum of probabilities for X=0 and X=1. It also explains how to find F(1.7) and F(4.5) by considering the highest discrete value less than or equal to the given 'x'.
The video illustrates how to write the CDF as a piecewise function, defining the probability for ranges of x values. It also shows how to graphically represent the CDF as a step function. The graph uses full circles to indicate included points (due to 'less than or equal to') and open circles for excluded points, demonstrating the cumulative nature of the probabilities.
The reverse process of deriving the PMF from a given CDF is explained. For any X_i (where i > 1), the PMF f(X_i) can be found by subtracting the CDF of the previous value (F(X_{i-1})) from the CDF of the current value (F(X_i)). For the first value X1, f(X1) is simply F(X1). An example demonstrates this method thoroughly by starting with a CDF and tabulating the corresponding PMF values.