This post has been de-listed
It is no longer included in search results and normal feeds (front page, hot posts, subreddit posts, etc). It remains visible only via the author's post history.
For example, I can estimate a linear regression for a dataset of 400k observations, and get the predicted values or the regressand in <3 seconds (using the statsmodels library)
However if I manually save all the coefficients from the regression into an array, and multiply each element of that array with the respective datapoint in the data matrix, and sum the resulting numbers for each row of the matrix (hence getting an identical array of predicted values as above), it's much longer. Using a nested for loop function it takes ~5 minutes, even using numba's @jit function.
What am I missing?
Subreddit
Post Details
- Posted
- 9 years ago
- Reddit URL
- View post on reddit.com
- External URL
- reddit.com/r/learnpython...