P. Andriotis, G. Oikonomou

Abstract:
Sentiment Analysis aims to extract information related to the emotional state of the person that produced a text document and also describe the sentiment polarity of the short or long message. This kind of information might be useful to a forensic analyst because it provides indications about the psychological state of the person under investigation at a given time. In this paper we use machine-learning algorithms to classify short texts (SMS), which could be found in the internal memory of a smartphone and extract the mood of the person that sent them. The basic goal of our method is to achieve low False Positive Rates. Moreover, we present two visualization schemes with the intention to provide the ability to digital forensic analysts to see graphical representations of the messaging activity of their suspects and therefore focus on specific areas of interest reducing their workload.
Reference:
P. Andriotis, G. Oikonomou, "Messaging Activity Reconstruction with Sentiment Polarity Identification", in Human Aspects of Information Security, Privacy, and Trust - HAS 2015, ser. Lecture Notes in Computer Science, 9190, pp. 475-486, 2015
Bibtex Entry:
@INPROCEEDINGS{Andriotis-2015-hcii,
  title = {Messaging Activity Reconstruction with Sentiment Polarity Identification},
  author = {Panagiotis Andriotis and George Oikonomou},
  publisher = {Springer},
  year = {2015},
  volume = {9190},
  editor = {Theo Tryfonas and Askoxylakis, {Ioannis G.}},
  series = {Lecture Notes in Computer Science},
  booktitle = {Human Aspects of Information Security, Privacy, and Trust - HAS 2015},
  pages = {475--486},
  doi = {10.1007/978-3-319-20376-8_42},
  gsid = {3416071116317224488},
  abstract = {Sentiment Analysis aims to extract information related to the emotional state of the person that produced a text document and also describe the sentiment polarity of the short or long message. This kind of information might be useful to a forensic analyst because it provides indications about the psychological state of the person under investigation at a given time. In this paper we use machine-learning algorithms to classify short texts (SMS), which could be found in the internal memory of a smartphone and extract the mood of the person that sent them. The basic goal of our method is to achieve low False Positive Rates. Moreover, we present two visualization schemes with the intention to provide the ability to digital forensic analysts to see graphical representations of the messaging activity of their suspects and therefore focus on specific areas of interest reducing their workload.},
}
Powered by bibtexbrowser

Messaging Activity Reconstruction with Sentiment Polarity Identification