Textual Analysis
- clearly much information is text based, and several
project topics are possible in this area also, some of
which are indicated below
Projects on machine translation of natural human languages
are possible, but may be somewhat dated, and pale in
comparison to the likes of Google translate with huge
processing power and databases using statistical methods.
However a niche, investigative problem within the area
could still provide a valid challenging project,
Authorship / Style verification
- adopts methods of pattern analysis to verify authorship
or stylistic approach
Applications clearly include analysis of ancient texts and
modern plagiarism detection!
Sentiment Analysis
- uses and extends a standard database of words or phrases,
and their conjunction, to determine whether the writer is
expressing positive, neutral or negative sentiments.
Applications include consumer and investor market research,
marketing and political campaigns.
Rule Extraction from text.
- attempts to formalise rules expressed in text
Applications generally include fairly restricted rule-based
applications such as law, standards and regulations.
A recent project included choosing valid module mix
fulfilling course regulations for undergraduate degrees.
Knowledge Extraction
More generally applicable to research areas where published
knowledge is extensive and increasing rapidly, so that
automated knowledge extraction methods are desirable if not
essential.
Applications include virtually every knowledge-based
endeavour known.