bab2min
commented
2 years ago Hi @Raymond-KT
As far as I know there are no Python tools to visualize hierarchical topic models.
So I recommend to extract the hierarchy of topics and its words from model and to draw the diagram manually.
You can print detail information of HLDA model using mdl.summary() method.
>>> mdl.summary(topic_word_top_n=5)
...
<Topics>
| #0 (15524, 25) : alabama war republican new nation
| #8 (17026, 20) : aristotl work use also first
| #16 (3211, 6) : centuri death greek bc wrote
| #19 (1168, 2) : award academi oscar ceremoni best
| #21 (102, 1) : book publish pictur children letter
| #22 (164, 1) : time tai second scale atom
| #23 (98, 1) : academi alain scienc mathemat geometri
| #26 (232, 1) : produc hollywood also motion pictur
| #27 (377, 1) : pari american work horn music
| #28 (175, 1) : astronom astronomi observ research physic
| #29 (1481, 1) : anthropolog cultur human social anthropologist
| #30 (1255, 1) : children social individu behavior studi
| #31 (1682, 1) : anarchist anarch movement feder worker
| #32 (1027, 1) : altruism love one studi good
| #33 (1103, 1) : rand novel work philosoph review
| #34 (747, 1) : galt one hank strike compani
| #9 (200, 1) : surfac solar area caus affect
| #17 (355, 1) : albedo effect snow reflect temperatur
| #10 (227, 1) : letter use alphabet form latin
| #18 (136, 1) : alpha script first phoenician variant
| #11 (245, 1) : soil plant develop field technolog
| #24 (380, 1) : agricultur scienc product crop anim
| #13 (71, 2) : film life group also environ
| #20 (75, 1) : maria catalan actress anna best
| #25 (155, 1) : alien song album seri game
|
Hello.
Thank you for making the package so that you can easily use various topic modeling techniques.
In the case of LDA, there is an example of visualizing. But, There seems to be no HLDA case.
Depending on your example code, I saved up to the tmm format file. But, I can't visualize the result.
I would like to ask if there is a way to visualize HLDA result so that Hierarchy can be seen.
Thank you.