@rickecon This is a follow-up on the question raised by @SixueLiuMACSS:
How are unsupervised learning and nonparametric statistics different?
Topics from the syllabi:
Nonparametric Inference (STAT 37400): kernel density estimation, local regression, splines, confidence bands, orthogonal functions, random processes, and sample methods...
Unsupervised learning (TTIC 31220): dimensionality reduction (PCA), clustering, topic modeling, HMM, density estimation, mixture models, and expectation maximization...
ML algorithms that are nonparametric
KNN (supervised)
Decision Trees (supervised)
SVM with RBF kernel (supervised)
KMeans (unsupervised)
PCA (unsupervised)
...
After a brief research on this issue, I find the possible answer is that the two fields overlap, but have different goals to achieve from the data. Also, the nonparametric approach to ML algorithms is getting more popular for its advantages in flexibility and good performance in prediction.
@rickecon This is a follow-up on the question raised by @SixueLiuMACSS:
How are unsupervised learning and nonparametric statistics different?
Unsupervised learning (TTIC 31220): dimensionality reduction (PCA), clustering, topic modeling, HMM, density estimation, mixture models, and expectation maximization...
Unsupervised Learning: find patterns, compress data, visualization...
After a brief research on this issue, I find the possible answer is that the two fields overlap, but have different goals to achieve from the data. Also, the nonparametric approach to ML algorithms is getting more popular for its advantages in flexibility and good performance in prediction.