Java代码实例

以下是一个简单的 Java 代码实例，展示了如何使用朴素贝叶斯算法进行文本分类：

import java.io.BufferedReader;
import java.io.FileReader;
import java.util.HashMap;
import java.util.Map;
import java.util.StringTokenizer;

public class NaiveBayesClassifier {

    private Map<String, Integer> wordCountMap = new HashMap<>();
    private Map<String, Double> wordProbMap = new HashMap<>();
    private int totalWords = 0;
    private int totalDocs = 0;

    public void train(String fileName) throws Exception {
        BufferedReader br = new BufferedReader(new FileReader(fileName));
        String line;
        while ((line = br.readLine()) != null) {
            StringTokenizer st = new StringTokenizer(line);
            String label = st.nextToken();
            while (st.hasMoreTokens()) {
                String word = st.nextToken();
                addWordToMap(word);
            }
            updateDocumentCount(label);
        }
        br.close();
        calculateWordProbabilities();
    }

    public String classify(String text) {
        double maxProb = Double.NEGATIVE_INFINITY;
        String bestLabel = "";
        StringTokenizer st = new StringTokenizer(text);
        while (st.hasMoreTokens()) {
            String word = st.nextToken();
            if (wordProbMap.containsKey(word)) {
                for (String label : wordCountMap.keySet()) {
                    double prob = Math.log(wordProbMap.get(word) / totalDocs);
                    prob += Math.log(wordCountMap.get(word + label) + 1);
                    prob -= Math.log(wordCountMap.get(label) + totalWords);
                    if (prob > maxProb) {
                        maxProb = prob;
                        bestLabel = label;
                    }
                }
            }
        }
        return bestLabel;
    }

    private void addWordToMap(String word) {
        if (wordCountMap.containsKey(word)) {
            wordCountMap.put(word, wordCountMap.get(word) + 1);
        } else {
            wordCountMap.put(word, 1);
        }
        totalWords++;
    }

    private void updateDocumentCount(String label) {
        if (wordCountMap.containsKey(label)) {
            wordCountMap.put(label, wordCountMap.get(label) + 1);
        } else {
            wordCountMap.put(label, 1);
        }
        totalDocs++;
    }

    private void calculateWordProbabilities() {
        for (String word : wordCountMap.keySet()) {
            for (String label : wordCountMap.keySet()) {
                double count = wordCountMap.containsKey(word + label) ? wordCountMap.get(word + label) : 0.0;
                double prob = (count + 1) / (wordCountMap.get(label) + totalWords);
                wordProbMap.put(word + label, prob);
            }
        }
    }

}

上述代码中，train() 方法用于训练分类器，classify() 方法用于对输入的文本进行分类。在训练阶段，程序会读取一个包含训练数据的文件，并统计单词出现的次数和文档的数量。在分类阶段，程序会遍历输入文本中的每个单词，并计算该单词在不同分类中出现的概率，然后选择概率最大的分类作为输出。