91Ó°ÊÓ

A stochastic matrix is an essential concept in linear algebra and probability theory. It's a square matrix used to describe the transitions of a Markov chain. Let’s break it down easily: a stochastic matrix is a matrix where each row represents a probability distribution.
Each entry of the matrix should be non-negative, meaning it cannot be less than zero. The entries are essentially probabilities showing the likelihood of moving from one state to another in a system.

• The matrix must have all non-negative entries.
• The sum of each row must equal 1.

In this context, each row of the matrix can be thought of as listing all the possible outcomes with their associated probabilities, summing up to a certainty (i.e., 1). Hence, each row can be seen as a complete and standalone probability distribution.

Non-negativity in matrices, particularly stochastic matrices, is a crucial requirement. In a non-negative matrix, no entries are below zero. This is because these matrices are often used to represent probabilities, which by definition cannot be negative.
For example, consider a matrix:\[ P = \begin{bmatrix} 0.2 & 1 \ 0.8 & 0 \end{bmatrix} \]
Each element such as 0.2, 1, 0.8, and 0 are non-negative, satisfying the first prerequisite of being a stochastic matrix.

• Every entry must be either zero or a positive number.
• Negative entries disrupt the probability interpretation of the matrix.

If just one entry turns negative, then the entire row fails to represent a plausible probability distribution, which is why this condition is so critical.

For a matrix to be considered stochastic, its rows need not only to be non-negative, but also each row must add up to 1. This property ensures that each row reflects a valid probability distribution.
Let's look at matrix \(P\):\[ P = \begin{bmatrix} 0.2 & 1 \ 0.8 & 0 \end{bmatrix} \]

• First row sum: \(0.2 + 1 = 1.2\)
• Second row sum: \(0.8 + 0 = 0.8\)

As seen above, neither the first row nor the second row sums to 1. This indicates that matrix \(P\) does not satisfy the row sum condition of a stochastic matrix.
When a row does not sum to 1, it implies that either there is an 'over' or 'under' probability in the system the matrix should represent, which defeats its purpose as a reliable probabilistic model. Therefore, ensuring that each row sums to 1 is central to maintaining the integrity of the stochastic matrix.