安卓计算句子余弦相似度

可以使用Java的自然语言处理库，如Stanford CoreNLP或LingPipe，来计算句子的余弦相似度。以下是一个使用Stanford CoreNLP计算余弦相似度的示例代码：

import edu.stanford.nlp.ling.Sentence;
import edu.stanford.nlp.linalg.ArrayMath;
import edu.stanford.nlp.pipeline.StanfordCoreNLP;
import edu.stanford.nlp.util.CoreMap;
import java.util.Arrays;
import java.util.List;
import java.util.Properties;

public class CosineSimilarity {

    public static double cosineSimilarity(String s1, String s2) {

        // Create StanfordCoreNLP object with NER and parse annotators
        Properties props = new Properties();
        props.setProperty("annotators", "tokenize, ssplit, parse");
        StanfordCoreNLP pipeline = new StanfordCoreNLP(props);

        // Tokenize and parse both sentences
        List<CoreMap> sentence1 = pipeline.process(Sentence.toWordList(s1)).get(CoreAnnotations.SentencesAnnotation.class);
        List<CoreMap> sentence2 = pipeline.process(Sentence.toWordList(s2)).get(CoreAnnotations.SentencesAnnotation.class);

        // Convert parsed sentences to word vectors
        double[][] vectors1 = sentence1.get(0).get(CoreAnnotations.SentencesAnnotation.class).get(0).get(SentenceVectorsAnnotation.class);
        double[][] vectors2 = sentence2.get(0).get(CoreAnnotations.SentencesAnnotation.class).get(0).get(SentenceVectorsAnnotation.class);

        // Compute cosine similarity using ArrayMath library
        double cosine = ArrayMath.cosine(vectors1[0], vectors2[0]);

        return cosine;
    }

    public static void main(String[] args) {

        String s1 = "This is a sentence.";
        String s2 = "This is another sentence.";

        double similarity = cosineSimilarity(s1, s2);

        System.out.println("Cosine similarity: " + similarity);
    }
}

另外，也可以使用第三方库，如Apache Commons Math，来计算余弦相似度。以下是一个使用Apache Commons Math计算余弦相似度的示例代码：

import org.apache.commons.math3.linear.ArrayRealVector;
import org.apache.commons.math3.linear.RealVector;

public class CosineSimilarity {

    public static double cosineSimilarity(String s1, String s2) {

        // Convert sentences to word vectors
        String[] words1 = s1.split(" ");
        String[] words2 = s2.split(" ");
        RealVector vector1 = new ArrayRealVector(words1.length);
        RealVector vector2 = new ArrayRealVector(words2.length);
        for (int i = 0; i < words1.length; i++) {
            vector1.setEntry(i, Double.parseDouble(words1[i]));
        }
        for (int i = 0; i < words2.length; i++) {
            vector2.setEntry(i, Double.parseDouble(words2[i]));
        }

        // Compute cosine similarity using dot product and magnitude
        double dotProduct = vector1.dotProduct(vector2);
        double magnitude = vector1.getNorm() * vector2.getNorm();
        double cosine = dotProduct / magnitude;

        return cosine;
    }

    public static void main(String[] args) {

        String s1 = "1 2 3";
        String s2 = "4 5 6";

        double similarity = cosineSimilarity(s1, s2);

        System.out.println("Cosine similarity: " + similarity);
    }
}