machine-learning - 记录实现朴素贝叶斯进行文本分类的可能性

Question

我正在实现朴素贝叶斯算法来进行文本分类。我有大约 1000 个用于培训的文档和 400 个用于测试的文档。我认为我已经正确实现了培训部分，但我在测试部分感到困惑。以下是我所做的简要说明:

在我的训练函数中:

vocabularySize= GetUniqueTermsInCollection();//get all unique terms in the entire collection

spamModelArray[vocabularySize]; 
nonspamModelArray[vocabularySize];

for each training_file{
        class = GetClassLabel(); // 0 for spam or 1 = non-spam
        document = GetDocumentID();

        counterTotalTrainingDocs ++;

        if(class == 0){
                counterTotalSpamTrainingDocs++;
        }

        for each term in document{
                freq = GetTermFrequency; // how many times this term appears in this document?
                id = GetTermID; // unique id of the term 

                if(class = 0){ //SPAM
                        spamModelArray[id]+= freq;
                        totalNumberofSpamWords++; // total number of terms marked as spam in the training docs
                }else{ // NON-SPAM
                        nonspamModelArray[id]+= freq;
                        totalNumberofNonSpamWords++; // total number of terms marked as non-spam in the training docs
                }
        }//for


        for i in vocabularySize{
                spamModelArray[i] = spamModelArray[i]/totalNumberofSpamWords;
                nonspamModelArray[i] = nonspamModelArray[i]/totalNumberofNonSpamWords;

        }//for


        priorProb = counterTotalSpamTrainingDocs/counterTotalTrainingDocs;// calculate prior probability of the spam documents
}

我认为我正确理解并实现了培训部分，但我不确定我是否可以正确实现测试部分。在这里，我尝试检查每个测试文档，并计算每个文档的 logP(spam|d) 和 logP(non-spam|d)。然后我比较这两个数量以确定类别(垃圾邮件/非垃圾邮件)。

在我的测试函数中:

vocabularySize= GetUniqueTermsInCollection;//get all unique terms in the entire collection
for each testing_file:
        document = getDocumentID;

        logProbabilityofSpam = 0;
        logProbabilityofNonSpam = 0;

        for each term in document{
                freq = GetTermFrequency; // how many times this term appears in this document?
                id = GetTermID; // unique id of the term 

                // logP(w1w2.. wn) = C(wj)∗logP(wj)
                logProbabilityofSpam+= freq*log(spamModelArray[id]);
                logProbabilityofNonSpam+= freq*log(nonspamModelArray[id]);
        }//for

        // Now I am calculating the probability of being spam for this document
        if (logProbabilityofNonSpam + log(1-priorProb) > logProbabilityofSpam +log(priorProb)) { // argmax[logP(i|ck) + logP(ck)]
                newclass = 1; //not spam
        }else{
                newclass = 0; // spam
        }

}//for

我的问题是；我想返回每个类别的概率，而不是精确的 1 和 0(垃圾邮件/非垃圾邮件)。我想看看例如newclass = 0.8684212 所以我可以稍后应用阈值。但我在这里很困惑。如何计算每个文档的概率？我可以使用 logProbabilities 来计算它吗？

Answer 1

由属于 < 类的一组特征 {F1, F2, ..., Fn} 描述的数据的概率em>C，根据朴素贝叶斯概率模型，为

P(C|F) = P(C) * (P(F1|C) * P(F2|C) * ... * P(Fn|C)) / P(F1, ..., Fn)

您拥有除 1/P( F1, ..., Fn) 之外的所有项(以对数形式)术语，因为您正在实现的朴素贝叶斯分类器中未使用该术语。 (严格来说，MAP 分类器。)

您还必须收集这些特征的频率，并从中计算

P(F1, ..., Fn) = P(F1) * ... * P(Fn)