- français
- English
Milestone 4
Metrics :
- Date a set of articles of the same year for KL and compute the error (Marc)
- Cosine similarity and date an article for distance1 and chi-square (Cynthia)
- New metric: out of place measure (Tao)
- Sentence length and ponctuation marks. (Malik, Gil)
Synonyms (Nicolas):
- Opt for the thesaurus database of synonyms, because the original one was not exhaustive and correct enough to produce interesting graphs. Adapt the SQL schema.
- Completion of the web interface, add a [0,1] normalization functionality and increase the data resolution (mean over 5 years instead of 10).
Topic Clustering : (Jeremy and Farah)
- Represent the article in an article-term matrix format.
- Apply Latent Dirichlet Allocation (LDA) to the matrix
- Ce wiki
- Cette page