Information Retrieval Techniques -- 2
Budget: $30 – $250 USD
I want to implement some information retrieval techniques in this dataset https://github.com/iMEdD-Lab/Greek_Parliament_Proceedings through a web based interface.
First task: We have to find the most important keywords and how they change over time, per speech, per member and for each party.
Second task :Given all the speeches, we need to detect similarities in pairs between members
of parliament. Specifically, we need to find a way to export a vector characteristics for each member
and then perform similarities(cosine) in pairs to detect the top-k pairs with the highest degree of similarity ( k is a parameter)
Third task:Considering all the speeches, we have to use the LSI technique in order to
find the most important subject areas and express each speech as a vector in a multidimensional space.
Optionally:We can implement some neural networks to extract some interesting relationship results.
First task: We have to find the most important keywords and how they change over time, per speech, per member and for each party.
Second task :Given all the speeches, we need to detect similarities in pairs between members
of parliament. Specifically, we need to find a way to export a vector characteristics for each member
and then perform similarities(cosine) in pairs to detect the top-k pairs with the highest degree of similarity ( k is a parameter)
Third task:Considering all the speeches, we have to use the LSI technique in order to
find the most important subject areas and express each speech as a vector in a multidimensional space.
Optionally:We can implement some neural networks to extract some interesting relationship results.