Stanford Algorithms Design and Analysis Part 2 week 2

最新推荐文章于 2019-04-05 12:25:00 发布

好大的房子

最新推荐文章于 2019-04-05 12:25:00 发布

阅读量228

点赞数

本文链接：https://blog.csdn.net/jffyuhgv/article/details/83868083

版权

分享一下我老师大神的人工智能教程！零基础，通俗易懂！http://blog.csdn.net/jiangjunshow

也欢迎大家转载本篇文章。分享知识，造福人民，实现我们中华民族伟大复兴！

Problem Set-2

Programming Assignment-2

Question 1

In this programming problem and the next you'll code up the clustering algorithm from lecture for computing a max-spacing -clustering. Download the text file here. This file describes a distance function (equivalently, a complete graph with edge costs). It has the following format:

[number_of_nodes]
[edge 1 node 1] [edge 1 node 2] [edge 1 cost]
[edge 2 node 1] [edge 2 node 2] [edge 2 cost]
...
There is one edge for each choice of , where is the number of nodes. For example, the third line of the file is "1 3 5250", indicating that the distance between nodes 1 and 3 (equivalently, the cost of the edge (1,3)) is 5250. You can assume that distances are positive, but you should NOT assume that they are distinct.

Your task in this problem is to run the clustering algorithm from lecture on this data set, where the target number of clusters is set to 4. What is the maximum spacing of a 4-clustering?

ADVICE: If you're not getting the correct answer, try debugging your algorithm using some small test cases. And then post them to the discussion forum!

import java.io.BufferedReader;import java.io.DataInputStream;import java.io.FileInputStream;import java.io.FileNotFoundException;import java.io.IOException;import java.io.InputStreamReader;import java.util.ArrayList;import java.util.Collections;/* * Question 1In this programming problem and the next you'll code up the clustering algorithm from lecture for computing a max-spacing k-clustering. Download the text file here. This file describes a distance function (equivalently, a complete graph with edge costs). It has the following format:[number_of_nodes][edge 1 node 1] [edge 1 node 2] [edge 1 cost][edge 2 node 1] [edge 2 node 2] [edge 2 cost]...There is one edge (i,j) for each choice of 1≤i<j≤n, where n is the number of nodes. For example, the third line of the file is "1 3 5250", indicating that the distance between nodes 1 and 3 (equivalently, the cost of the edge (1,3)) is 5250. You can assume that distances are positive, but you should NOT assume that they are distinct.Your task in this problem is to run the clustering algorithm from lecture on this data set, where the target number k of clusters is set to 4. What is the maximum spacing of a 4-clustering?ADVICE: If you're not getting the correct answer, try debugging your algorithm using some small test cases. And then post them to the discussion forum! */public class PS2Q1 { static int[] parents; /* parents[i] if -ve it is parent to itself. and the value represents its count- size of the cluster. */ static class Edge implements Comparable<Edge>{  int i;  int j;  int cost;  public Edge(int i, int j, int cost){   this.i = i;   this.j = j;   this.cost = cost;  }  @Override  public int compareTo(Edge arg0) {   // TODO check the order   final int BEFORE = 1;   final int AFTER = -1;   if (this.cost >= arg0.cost) return BEFORE;   else return AFTER;   //return 0;  } } public static int find(int i){  while (parents[i]>0){   i = parents[i];  }  return i; } public static void union(int i, int j){  //find parents of i and j..if same..they r in same cluster, update the count  //else in diff clusters..choose the one with less count and change its parent to the other parent  int pi = find(i);  int pj = find(j);  if (pi == pj) parents[pi] += -1;  else {   if (parents[pi] < parents[pj]){    //-ve counts so actually pi is larger than pj    parents[pi] += parents[pj];    parents[pj] = pi;   }else{    parents[pj] += parents[pi];    parents[pi] = pj;   }  }   } public static int numClusters(){  //num of -ve entries in parents is the no of clusters  int c = 0;  for(int i = 0; i < parents.length; i++)   if (parents[i]<0) c++;  return c;   }  /**  * @param args  */ public static void main(String[] args) {  ArrayList<Edge> edges = new ArrayList<Edge>();  ArrayList<ArrayList<Integer>> clusters = new ArrayList<ArrayList<Integer>>();  int k = 4;  int n = 0;  try {   FileInputStream f = new FileInputStream(".//Algo2//clustering1.txt");   DataInputStream d = new DataInputStream(f);   BufferedReader b =  new BufferedReader(new InputStreamReader(d));   n = Integer.parseInt(b.readLine());   parents = new int[n];      for(int i = 0; i < n;i++){    parents[i] = -1;       }   String str; int i,j,v;   while((str=b.readLine())!=null){    i = Integer.parseInt(str.split(" ")[0]);    j = Integer.parseInt(str.split(" ")[1]);    v = Integer.parseInt(str.split(" ")[2]);    edges.add(new Edge(i-1,j-1,v));   }   Collections.sort(edges);   for(Edge e : edges){    union(e.i,e.j);    if (numClusters()==k) {     System.out.println("k clusters found ");     break;        }   }   //print the max distance among clusters..is actuallz the min distance between any two clusters   int max = Integer.MAX_VALUE;   for (Edge e : edges){    if (find(e.i)!=find(e.j)) max=Math.min(max, e.cost);       }   System.out.println("max distance "+max);  } catch (FileNotFoundException e) {   // TODO Auto-generated catch block   e.printStackTrace();  } catch (NumberFormatException e) {   // TODO Auto-generated catch block   e.printStackTrace();  } catch (IOException e) {   // TODO Auto-generated catch block   e.printStackTrace();  }   //create n clusters  for(int i = 0; i <n;i++)   clusters.add(new ArrayList<Integer>(i));  while(clusters.size() > k){  } }}

Question 2

In this question your task is again to run the clustering algorithm from lecture, but on a MUCH bigger graph. So big, in fact, that the distances (i.e., edge costs) are only defined implicitly, rather than being provided as an explicit list.

The data set is here. The format is:
[# of nodes] [# of bits for each node's label]
[first bit of node 1] ... [last bit of node 1]
[first bit of node 2] ... [last bit of node 2]
...
For example, the third line of the file "0 1 1 0 0 1 1 0 0 1 0 1 1 1 1 1 1 0 1 0 1 1 0 1" denotes the 24 bits associated with node #2.

The distance between two nodes and in this problem is defined as the Hamming distance--- the number of differing bits --- between the two nodes' labels. For example, the Hamming distance between the 24-bit label of node #2 above and the label "0 1 0 0 0 1 0 0 0 1 0 1 1 1 1 1 1 0 1 0 0 1 0 1" is 3 (since they differ in the 3rd, 7th, and 21st bits).

The question is: what is the largest value of such that there is a -clustering with spacing at least 3? That is, how many clusters are needed to ensure that no pair of nodes with all but 2 bits in common get split into different clusters?

NOTE: The graph implicitly defined by the data file is so big that you probably can't write it out explicitly, let alone sort the edges by cost. So you will have to be a little creative to complete this part of the question. For example, is there some way you can identify the smallest distances without explicitly looking at every pair of nodes?

import java.io.BufferedReader;import java.io.DataInputStream;import java.io.FileInputStream;import java.io.FileNotFoundException;import java.io.IOException;import java.io.InputStreamReader;import java.util.ArrayList;import java.util.BitSet;import java.util.HashMap;import java.util.Map.Entry;/* * In this question your task is again to run the clustering algorithm from lecture, but on a MUCH bigger graph. So big, in fact,  * that the distances (i.e., edge costs) are only defined implicitly, rather than being provided as an explicit list.The data set is here. The format is:[# of nodes] [# of bits for each node's label][first bit of node 1] ... [last bit of node 1][first bit of node 2] ... [last bit of node 2]...For example, the third line of the file "0 1 1 0 0 1 1 0 0 1 0 1 1 1 1 1 1 0 1 0 1 1 0 1" denotes the 24 bits associated with node #2.The distance between two nodes u and v in this problem is defined as the Hamming distance--- the number of differing bits --- between the two nodes' labels. For example, the Hamming distance between the 24-bit label of node #2 above and the label "0 1 0 0 0 1 0 0 0 1 0 1 1 1 1 1 1 0 1 0 0 1 0 1" is 3 (since they differ in the 3rd, 7th, and 21st bits).The question is: what is the largest value of k such that there is a k-clustering with spacing at least 3? That is, how many clusters are needed to ensure that no pair of nodes with all but 2 bits in common get split into different clusters?NOTE: The graph implicitly defined by the data file is so big that you probably can't write it out explicitly, let alone sort the edges by cost. So you will have to be a little creative to complete this part of the question. For example, is there some way you can identify the smallest distances without explicitly looking at every pair of nodes? *  */public class PS2Q2 {  /*  * This problem is exactly like PS2Q1 except that there are too many comparisons to make if we go with crude approach.   * 20000 x 20000 comparisons needed  */ //point and its leader static HashMap<BitSet, BitSet> clusters = new HashMap<BitSet, BitSet>(); static int n; static int numBits; public static BitSet getBitSet(String str){  String str2[] = str.split(" ");  BitSet b = new BitSet(numBits);  int j = numBits-1;  b.clear();  for (int i = 0; i < str2.length; i++){   if (Integer.parseInt(str2[i]) == 1){    b.flip(j);   }   j--;  }  return b; } public static BitSet find(BitSet b){  while (!b.equals(clusters.get(b))){   //b = (BitSet) clusters.get(b).clone();   b = clusters.get(b);  }  return b; } public static void union (BitSet a, BitSet b){  //actually smaller cluster should be merged with bigger one. Here do it randomly. Cluster sizes should be maintained for  //it to work.  BitSet pa = find(a);  BitSet pb = find(b);  if (!pa.equals(pb)){   clusters.put(pa, pb);  } } public static ArrayList<BitSet> getMembers(BitSet s){  BitSet sbackup = (BitSet) s.clone();  ArrayList<BitSet> ret = new ArrayList<BitSet>();  for(int i = 0; i <= numBits-1; i++){   BitSet s1 = new BitSet();   s1.clear();   s1 = (BitSet) sbackup.clone();   s1.flip(i);   if (clusters.containsKey(s1)){    ret.add(s1);   }  }  //now flip 2 bits to create distance of 2  for(int i = 0; i <= numBits-1; i++){   BitSet s1 = new BitSet();   s1.clear();   s1 = (BitSet) sbackup.clone();   s1.flip(i);   for (int j = i+1; j<=numBits-1; j++){    BitSet s2 = new BitSet();    s2 = (BitSet) s1.clone();    s2.flip(j);    if (clusters.containsKey(s2)) ret.add(s2);   }  }  return ret; } public static void main(String[] args) {  //int distance = 2; //must be <= 2  try {   FileInputStream f = new FileInputStream(".//Algo2//clustering2.txt");   DataInputStream d = new DataInputStream(f);   BufferedReader br = new BufferedReader(new InputStreamReader(d));   String str = br.readLine();   n = Integer.parseInt(str.split(" ")[0]);   numBits = Integer.parseInt(str.split(" ")[1]);   int count2 = 0;   while((str = br.readLine())!= null){    BitSet b = getBitSet(str);//    if (clusters.containsKey(b)) {//     System.out.println(" a duplicate found " + b.toString());//    }    clusters.put(b, b);    //count2++;   }   //System.out.println( count2 + " entries are read ");   //System.out.println(" number of entries in DHT " + clusters.size());   for (BitSet s : clusters.keySet()){    //for all at distance of 1 or 2 from s    ArrayList<BitSet> members = getMembers(s);    //System.out.println(" members sizes "+members.size());    if (members.size() == 0) count2++;    for (BitSet m : members){     union(s,m);    }   }   System.out.println(" number of points with zero neighbours with <=2 distance "+count2);   int count = 0;   //parent of a parent is itself..each cluster has a single parent.    for(Entry<BitSet, BitSet> e : clusters.entrySet()){    if (e.getKey().equals(e.getValue())){     count++;    }   }   System.out.println(" num clusters " + count);  } catch (FileNotFoundException e) {   // TODO Auto-generated catch block   e.printStackTrace();  } catch (IOException e) {   // TODO Auto-generated catch block   e.printStackTrace();  } }}