[New features]
- We add sentiment analysis, which is useful for determining the emotional tone of a piece of text, such as a customer review or social media comment.
- We add WorcCloud, which is a simple yet powerful way to see which words appear most often in a collection of text.
Thanks to @tlehao who worked for these
- We add Shifterator, which is useful for visualizing pairwise comparisons between texts through word shifts.
[Interface]
Some adjustments are needed to make it easier to move from one menu to another.
[Data Source]
- Now supports PubMed Central Article Datasets (.tar.gz or .xml). Please visit PMC.
- Now supports HathiTrust's JSON file. Please visit HathiTrust.
- Now supports Dimension's CSV file. Please visit Dimension.
Thanks to @tlehao who worked for these
- Now supports OpenAlex's CSV file. Please visit OpenAlex.
- Now supports XLS and XLSX files.
[Topic Modeling]
- Enables fine-tuning of topic representations using a KeyBERT-like model, supported by large language models such as OpenAI’s GPT-4o or Google’s Flan-T5.
Thanks to @tlehao who worked for this
- New option in pyLDA that allows you to adjust the Lambda, and you can now download the results as a file.
[Burst Detection]
Fix the issue on the visualization and now support n-grams.
[Sunburst]
Support filter by citation or certain text and focus on a specific source.
Thanks to @tlehao who worked for this
Contributors
tlehao