The calculation of
PageRank is to find the inherent vector of the maximum characteristic value of the ranks of the passing probabilities.
this is because, when t approaches infinity asymptotically linear transformation, we can transform according to the ranks of the " the maximum value of the absolute value of " and " characteristics; natural vector " belong to it; it will be fundamentally put down. In other words, the probability process with the rank of the passing probability is a process of multiplication and multiplication of the column, and the probability of the forward state can be calculated.
furthermore, although it may be difficult to find, the value of the intrinsic value and the intrinsic vector is a basic mathematical tool that can be closely analyzed. We are free to assign the initial value of the vector, but because the columns are multiplied, the resulting vector will be concentrated in a particular set of values. We combine those stable values is called natural vector, scalar characteristics inherent in a vector (scalar) is called the characteristic value, the general calculation method for such decomposition characteristics, the solution of characteristic value problem is called the characteristic value problem.
(* Note) for N times Affirmative ranks A to meet the number of lambda Ax = lambda x called A values, X is called a natural vector belongs to a. If you can not adapt to the concept of the ranks, you can also consider N× N two yuan arrangement on it. At the same time, you can also consider the vector into a length of N of the ordinary (one dollar) arrangement on it.
simple example of
page links over each other
first, the adjacency list this passage diagram structure represented as order type, there are the following. That is, according to the source of each link ID lists the target ID.
I D link link source; 2,3 />1 ID