LSA

[@shreysoni4107]

Very easy explanation! Thank you for making it simple.

[@sujits3458]

Very good explanation, I was able to understand the concept so easily, thank you :-)

[@dharmendrathakur1487]

Thank You very much Francisco...You made it so simple :-)

[@pratikanand2200]

very very good description. can you upload the link of this ppt?

[@kkhanh4y]

rows represent documents and columns represent words, isn't it? but here rows represent words and columns represent documents.

[@fiacobelli]

You can use different matrices to judge word similarity or document similarity. In the end it is the same. All that changes is what vector you use depending of your end goal.

[@kkhanh4y]

@@fiacobelli Ok, thank you, dear

[@pratikanand2200]

the meaning of a word is not satisfying as the passages contain other words also so can the average meaning is given to all words that the passages contain BUT the LSA describe it properly.

[@2441139knakmg]

Best

[@mohitpahuja6894]

wrong explanation. you havent actually reduced dimensionalty. what you have found is approximation of data. what you need to do is get projection of data on top two eigenvectors to get lower dimensional data

[@fiacobelli]

Correct. It is an approximation matrix. You can play with dimensionality as well, but this is the seminal concept behind LSA

[@harryshuman9637]

COMP 596 is wack