Untitled

作者: TongWei | 来源:发表于2014-09-03 01:49 被阅读0次

Step 1

Input file type : Sequence File Format

Each line o input file:

Json parser: jackson

Pange Rank

PageRank Equation:

Map

Input : <nid n, Node N>  

Output :<nid n, Node N> <nid AdjId, value mass>

Reduce

Input: <nid n, [Node N, value massi]>

Output: <nid n, Node N>   N.pr = sum(massi);

Mass Loss at Dangling Nodes: Redistributed  mass across all nodes evenly.

Redistribution process: m (mass loss), no reducers.

One iteration of PageRank = (Map + Reduce) + Map

After each iteration: check if converge.

Rank stability VS Value Convergence

相关文章

网友评论

      本文标题:Untitled

      本文链接:https://www.haomeiwen.com/subject/poyytttx.html