Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Ensemble Topic Modelling

4c76f001e0a3d59cc5a269df70940dfd?s=47 Leland McInnes
July 12, 2019
Research
1
210

Ensemble Topic Modelling

A short lightning talk on ensemble topic modelling with pLSA using the enstop package.

4c76f001e0a3d59cc5a269df70940dfd?s=128

Leland McInnes

July 12, 2019
Tweet

More Decks by Leland McInnes

See All by Leland McInnes
4c76f001e0a3d59cc5a269df70940dfd?s=48 lmcinnes
0
230
4c76f001e0a3d59cc5a269df70940dfd?s=48 lmcinnes
0
14
4c76f001e0a3d59cc5a269df70940dfd?s=48 lmcinnes
1
98
4c76f001e0a3d59cc5a269df70940dfd?s=48 lmcinnes
2
2.8k
4c76f001e0a3d59cc5a269df70940dfd?s=48 lmcinnes
3
960
4c76f001e0a3d59cc5a269df70940dfd?s=48 lmcinnes
2
1.4k

Other Decks in Research

See All in Research
D9a4e35302e14ce9d8db3474ca4d129d?s=48 yutaogura
0
210
557c7fa1fbf43b88830901feb21f95a0?s=48 ynakano
1
130
14da6ebc2e909305afdb348e7970de81?s=48 wingnus
1
1.2k
6027044e52fbcbc07f22d0235fc9da37?s=48 secondapunta
0
140
0d4ef9af8e4f0cf5c162b48ba24faea6?s=48 ept
0
12k
8a7e83d2e447783ab6d824f553429a09?s=48 stakaya
2
230
31bf3df4bac9b584c89bc11b0c84e132?s=48 resnant
1
1.9k
C612ab39597a17ba5948cae54d13f99f?s=48 mkimura
3
3.2k
C31b5afbe1ffb5588b2fe5a9e39b357b?s=48 takarasawa_
1
490
65241d0356dc5a93b0fec68bcf7b42ac?s=48 yumulab
0
7.1k
89e5ca0f52e56459bde438758eca64d2?s=48 expertgroup
0
330
0f2a473c07602f3dd53c5ed0de0c56b5?s=48 danielkatz
PRO
0
2.6k

Featured

See All Featured
7cbd3f5f922d798cc312eaf596de7ae6?s=48 iamctodd
28
2.6k
A9a491b0fcbe0fbce3d64063a37add99?s=48 rmw
15
990
78b475797a14c84799063c7cd073962f?s=48 holman
450
140k
88bc30284c6a424b63a92aad27d0ba36?s=48 jnunemaker
PRO
43
4.8k
D8fc2580cfaca035f666d9e4ee79a7f7?s=48 mongodb
24
4k
Bfd6b681ec6e8c9bef82ff0521c364f7?s=48 kastner
55
2k
C0b42577cfc2b321be3474618107e933?s=48 samanthasiow
61
6.6k
C157b16234f1e75e8eac3698c1d4414a?s=48 ddemaree
273
32k
C4de88ea47538e4594750e3861a341c8?s=48 brettharned
96
3.2k
2f5463832ccb768ccb4a1ca3607c27ef?s=48 jacobian
260
20k
D36d2c1821af9249b69ff7f5ed60529b?s=48 tammielis
239
23k
6804f1775cb4babfcc3851298566fbce?s=48 tanoku
91
8.8k

Transcript

  1. Ensemble Topic Modelling Leland McInnes

  2. Model a corpus of documents in terms of underlying “topics”

  3. Topic Modelling as Matrix Factorization

  4. None
  5. None
  6. None
  7. None

  8. LDA and pLSA are probabilistic matrix factorization methods

  9. (Ensembles of) pLSA

  10. Performance?

  11. None

  12. Quality?

  13. None

  14. Instability?

  15. These are hard optimization problems

  16. Topics vary from one run to another

  17. What are the stable topics? Inspired by https://github.com/RaRe-Technologies/gensim/pull/2282

  18. None

  19. Each cluster represents a stable topic

  20. None

  21. • Greater stability • Determines number of topics automatically •

    Embarrassingly parallel computation

  22. Implementation

  23. sklearn API

  24. None

  25. https://github.com/lmcinnes/enstop

  26. pip install enstop