final Node node(int weight, int count, Node tail) {
Node result = nodes[nextNode++];
result.weight = weight;
result.count = count;
result.tail = tail;
return result;
}
/**
* This is exactly as Head Insertion of a Linked List. We choose Head Insertion to get an O(1)
* Insertion into an Adjacency List, Tail Insertion can also be used
*
* @param head
* @param value
* @return
*/
private static Node insert(Node head, int value, int weight) {
Node p = new DijsktraAlgo().new Node();
p.value = value;
p.next = head;
p.visited = false;
p.weight = weight;
return p;
}
/**
* the array weights should have a weight for each document ID, that is weights[docID] is the
* weight for docID
*
* @param weights
*/
public void genericWeight(double[] weights) {
Node temp = head;
while (temp != null) {
temp.weight *= weights[temp.docID];
temp = temp.next();
}
}
/** Apply log frequency normalization to this postings list */
public void logNormalizeTermFrequencies() {
Node temp = head;
while (temp != null) {
temp.weight = 1 + Math.log(temp.weight);
temp = temp.next();
}
}
/** multiply all of the weights by log(number_of_docs/number_of_docs_with_this_term) */
public void idfWeight(int numDocs) {
double idfWeight = Math.log((double) numDocs / occurrences);
Node temp = head;
while (temp != null) {
temp.weight *= idfWeight;
temp = temp.next();
}
}
/** Apply boolean term normalization to this postings list */
public void booleanNormalizeTermFrequencies() {
Node temp = head;
while (temp != null) {
if (temp.weight > 0) {
temp.weight = 1;
}
temp = temp.next();
}
}
/**
* Since this is for boolean queries, the score should be 1.0 for all documents that are in this
* posting list
*/
public double[] getScores() {
if (head == null) {
return null;
} else {
double[] scores = new double[occurrences];
Node temp = head;
int i = 0;
while (temp != null) {
scores[i] = temp.weight();
temp = temp.next();
i++;
}
return scores;
}
}
