3 ms·
I was a fan of LDA but would not agree that it is 'probably useful' today. It's an unsupervised clustering algorithm based on Gibbs sampling. Like k-means, it's
by bpiche 2y ago
I was a fan of LDA but would not agree that it is 'probably useful' today. It's an unsupervised clustering algorithm based on Gibbs sampling. Like k-means, it's gonna return a few buckets that will have to be reviewed by a human for data exploration. In this case instead of neatly labeled buckets, these are unlabeled distributions of distributions (lists of single word tokens). If you do some kind of multiword tokenization preprocessing, it'll return a few lists of words and multiword tokens for each document. How is this useful to an end user? Even internally, they're not useful embeddings/vectorizations. Would love to hear some contrary opinions
- choppaface 2y agoIn many applications like especially Google's display ad targeting market, the "accuracy" of the clusters isn't so import as the lift in key metrics (e.g. click rates or revenue) and the overall efficiency of the method. Indeed the clustering algo might get things "dead wrong" but somehow surface something that causes clicks and revenue to increase. LDA offered much improvement over e.g. TF-IDF models, just as t-SNE improved on LDA, and now LLM embeddings are on average better and potentially cheap to compute. LDA could be useful if your success metric is perplexity; k-means is useful if vector distance is very meaningful for your problem. Also well-studied algorithms are generally useful for initial studies in a new, unknown dataset. As always with ML, the dataset and setting are just as important as the model and algorithm.
- bpiche 2y agoThank you for the well considered response