search

AntigoniKourou / Thesis

4 stars 1 forks source link

tasks for april 19 #21

Open maartenmarx opened 8 years ago

maartenmarx commented 8 years ago
  1. Research questions, worked out very well
  2. Scan of your pipeline
  3. Data: amsterdam: all preprocessing ready (nltk or pattern (has sentiment scorer). )
  4. Make your ontology (list of features/aspects).
  5. Start on splitting the reviews based on your ontology (every aspect is a colored marker: mark the reviews with your markers(the snippet/piece of text belonging to that aspect).
  6. Present statistics about all these data
AntigoniKourou commented 8 years ago

  1. Defined the RQ as follows (full text can be found here or latex):

    Firstly, how good can feature-based opinion mining estimate the quality of customer requirements represented by accommodation features in e-tourism feedback systems? Secondly, how will the implementation of this pipeline affect the customer satisfaction during the decision making process?

  2. Scanned and figure
  3. All pre-processing done. Script includes: language check, tokenize, sentence split, lemmatize -- The data is from all NL not Amsterdam only
  4. Accommodation ontology created, as well as script for finding the lemmas of all synonyms/hyponyms/hypernyms of the ontology features, based on human logic.
  5. Not done. At the moment working at it.