An e-mail analysis method based on text mining techniques

The impact of indexing approaches on Arabic text classification

“…Automatic Text Classification (ATC) is a machine learning technique which aims at assigning automatically a text document to a thematic category from a predefined set of classes. ATC has an important role in many information retrieval applications: website classification [1–4], automatic indexing [5], email filtering [6, 7], spam filtering [8–10], ontology mapping [11], hypertext classification [12], and sentiment analysis [13]. Many statistical and supervised learning algorithms have been applied to text categorization: naïve Bayes [14], k-nearest neighbor [15], support vector machines [16], decision tree [17], maximum entropy [18], hidden Markov model [19], and neural network [20].…”

Improving the performance of Naive Bayes multinomial in e-mail foldering by introducing distribution-based balance of datasets

“…In this study, we deal with a particular task within text classification: classification of e-mail into predefined folders (e-mail foldering Bekkerman, McCallum, & Huang, 2005;Klimt & Yang, 2004). E-mail classification has been studied for example in the classification or filtering of junk mail (or spam Guzella & Caminhas, 2009;Sahami, Dumais, Heckerman, & Horvitz, 1998) and in the analysis of e-mails collected at a customer center (Sakurai & Suyama, 2005), but its application to (semi) automatic classification of incoming mail into folders defined by user has not received so much attention.…”

Text Mining Infrastructure inR

“…Further common scenarios are help desk inquiries (Sakurai and Suyama 2005), measuring customer preferences by analyzing qualitative interviews (Feinerer and Wild 2007), automatic grading (Wu and Chen 2005), fraud detection by investigating notification of claims, or parsing social network sites for specific patterns such as ideas for new products.…”