问题
I have a LDA model with the 10 most common topics in 10K documents. Now it's just an overview of the words with corresponding probability distribution for each topic.
I was wondering if there is something available for python to visualize these topics?
回答1:
pyLDAvis looks reasonably good.
There's also Termite developed by Jason Chuang of Stanford.
回答2:
There some visulizations you can choise.
In the topic of Visualizing topic models, the visualization could be implemented with,
- D3 and Django(Python Web), e.g. Circle Packing, or Site Tag Explorer, etc
- Network X
In this topic Visualizing Topic Models, the visualization could be implemented with
- Matplotlib
- Bokeh
- etc.
回答3:
Word clouds are popular ways of visualizing topic distributions. To generate a word cloud in python consider cloning the wordcloud library.
回答4:
Depending upon the use case, you may chose some of these Visualizations as well.
1)Stacked Bar Chart
2)Heatmap
3)Table (https://de.dariah.eu/tatom/topic_model_visualization.html)
4)Histogram
5)Violin Plot (https://www.analyticsvidhya.com/blog/2015/05/data-visualization-python/)
Most of these visualizations are possible thorough Matplotlib library in Python.
回答5:
PyLDAVis - Visualization tool for LDA Models
Useful article for learning Topic Modelling using different models in Python
来源：https://stackoverflow.com/questions/30397550/visualizing-an-lda-model-using-python