Question

我试图在以下步骤中找出句子是正面还是负面：

1。）使用斯坦福NLP解析器从句子中检索词性（动词，名词，形容词等）。

2。）使用 SentiWordNet 查找与每个词性相关的正值和负值。

3。）总结所获得的正值和负值，以计算与句子相关的净正和净负值值。

但问题在于，SentiWordNet根据不同的感官/背景返回正/负值列表。是否可以将特定句子与词性一起传递给SentiWordNet解析器，以便它可以自动判断sense / context并返回只有一对的正值和负值？

或者还有其他替代解决方案吗？

感谢。

Answer 1

SentoWordNet Demo Code 这可能会对你有帮助。

//    Copyright 2013 Petter Törnberg
//
//    This demo code has been kindly provided by Petter Törnberg <pettert@chalmers.se>
//    for the SentiWordNet website.
//
//    This program is free software: you can redistribute it and/or modify
//    it under the terms of the GNU General Public License as published by
//    the Free Software Foundation, either version 3 of the License, or
//    (at your option) any later version.
//
//    This program is distributed in the hope that it will be useful,
//    but WITHOUT ANY WARRANTY; without even the implied warranty of
//    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
//    GNU General Public License for more details.
//
//    You should have received a copy of the GNU General Public License
//    along with this program.  If not, see <http://www.gnu.org/licenses/>.

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

public class SentiWordNetDemoCode {

    private Map<String, Double> dictionary;

    public SentiWordNetDemoCode(String pathToSWN) throws IOException {
        // This is our main dictionary representation
        dictionary = new HashMap<String, Double>();

        // From String to list of doubles.
        HashMap<String, HashMap<Integer, Double>> tempDictionary = new HashMap<String, HashMap<Integer, Double>>();

        BufferedReader csv = null;
        try {
            csv = new BufferedReader(new FileReader(pathToSWN));
            int lineNumber = 0;

            String line;
            while ((line = csv.readLine()) != null) {
                lineNumber++;

                // If it's a comment, skip this line.
                if (!line.trim().startsWith("#")) {
                    // We use tab separation
                    String[] data = line.split("\t");
                    String wordTypeMarker = data[0];

                    // Example line:
                    // POS ID PosS NegS SynsetTerm#sensenumber Desc
                    // a 00009618 0.5 0.25 spartan#4 austere#3 ascetical#2
                    // ascetic#2 practicing great self-denial;...etc

                    // Is it a valid line? Otherwise, through exception.
                    if (data.length != 6) {
                        throw new IllegalArgumentException(
                                "Incorrect tabulation format in file, line: "
                                        + lineNumber);
                    }

                    // Calculate synset score as score = PosS - NegS
                    Double synsetScore = Double.parseDouble(data[2])
                            - Double.parseDouble(data[3]);

                    // Get all Synset terms
                    String[] synTermsSplit = data[4].split(" ");

                    // Go through all terms of current synset.
                    for (String synTermSplit : synTermsSplit) {
                        // Get synterm and synterm rank
                        String[] synTermAndRank = synTermSplit.split("#");
                        String synTerm = synTermAndRank[0] + "#"
                                + wordTypeMarker;

                        int synTermRank = Integer.parseInt(synTermAndRank[1]);
                        // What we get here is a map of the type:
                        // term -> {score of synset#1, score of synset#2...}

                        // Add map to term if it doesn't have one
                        if (!tempDictionary.containsKey(synTerm)) {
                            tempDictionary.put(synTerm,
                                    new HashMap<Integer, Double>());
                        }

                        // Add synset link to synterm
                        tempDictionary.get(synTerm).put(synTermRank,
                                synsetScore);
                    }
                }
            }

            // Go through all the terms.
            for (Map.Entry<String, HashMap<Integer, Double>> entry : tempDictionary
                    .entrySet()) {
                String word = entry.getKey();
                Map<Integer, Double> synSetScoreMap = entry.getValue();

                // Calculate weighted average. Weigh the synsets according to
                // their rank.
                // Score= 1/2*first + 1/3*second + 1/4*third ..... etc.
                // Sum = 1/1 + 1/2 + 1/3 ...
                double score = 0.0;
                double sum = 0.0;
                for (Map.Entry<Integer, Double> setScore : synSetScoreMap
                        .entrySet()) {
                    score += setScore.getValue() / (double) setScore.getKey();
                    sum += 1.0 / (double) setScore.getKey();
                }
                score /= sum;

                dictionary.put(word, score);
            }
        } catch (Exception e) {
            e.printStackTrace();
        } finally {
            if (csv != null) {
                csv.close();
            }
        }
    }

    public double extract(String word, String pos) {
        return dictionary.get(word + "#" + pos);
    }

    public static void main(String [] args) throws IOException {
        if(args.length<1) {
            System.err.println("Usage: java SentiWordNetDemoCode <pathToSentiWordNetFile>");
            return;
        }

        String pathToSWN = args[0];
        SentiWordNetDemoCode sentiwordnet = new SentiWordNetDemoCode(pathToSWN);

        System.out.println("good#a "+sentiwordnet.extract("good", "a"));
        System.out.println("bad#a "+sentiwordnet.extract("bad", "a"));
        System.out.println("blue#a "+sentiwordnet.extract("blue", "a"));
        System.out.println("blue#n "+sentiwordnet.extract("blue", "n"));
    }
}

Answer 2

我们可以将pos传递给sentiwordnet解析器。下载模式python模块

from pattern.en import wordnet

print wordnet.synsets("kill",pos="VB")[0].weight

wordnet.synsets返回同义词列表从那我们选择第一项输出将是（极性，主观性）的元组希望这会有所帮助...

情绪分析（SentiWordNet） - 判断句子的上下文

2 个答案: