I was trying to simulate multivariate normal data in python the following way However, the resulting variables are always perfectly linearly correlated. I then checked np.linalg.eigvals(Sigma) and realized that Sigma is not positive semi-definite. This surprised me since 1) I did not get an error from multivariate_normal and 2) Sigma was generated from an outer product, which is supposed to
Tag: numpy
How to split a 2d numpy array vertically into a new 2d numpy array?
I have this code that essentially splits a two-dimensional numpy array horizontally and makes a new two-dimensional numpy array out of it. Output of my code: How can I do this with less lines of code? I assume it could be very resource intensive, as soon as I apply this example to my larger task. Answer I suppose using numpy.hsplit
Python numba returned data types when calculating MSE
I am using numba to calculate MSE. The input are images which are ready as numpy arrays of uint8. Each element is 0-255. When calculating the squared difference between two images the python function returns (expectedly) a uint8 result, but the same function when using numba returns int64. What’s unclear to me is why the python-only code preserves the data-type
How to inverse the rgb image color from “white-black” to “black-white” in matplotlib
The following two images are from the Mnist dataset. The image_7 is a gray image with shape (28, 28) while image_2 is (28, 28, 3). I want to inverse the image colors and display it in the matplotlib. As you can see, the second image is inversed. But how to inverse the fourth image from “white-black” to “black-white”. Answer A
How to get prior close when you have all stocks in a single DF?
Sorry for the noob question. I have a bunch of stocks in a sqlite3 database: When I print the df, it gives me the following (where each stock_id refers to a unique stock, e.g APPL): I need to target each unique stock_id individually, and get the prior close. I know if each stock was in its own separate dataframe, I
fastest way to reshape 2D numpy array (gray image) into a 3D stacked array
I have a 2D image with the shape (M, N), and would like to reshape it into (M//m * N//n, m, n). That is, to stack small patches of images into a 3D array. Currently, I used two for-loop to achieve that Is there any other faster way to do this? Thanks a lot! Answer Use skimage.util.view_as_blocks: Output: NB. Be
Alternating column values
I am working on a project where my dataset looks like bellow: Origin Destination Num_Trips Hamburg Frankfurt 2 Hamburg Cologne 1 Cologne Hamburg 3 Frankfurt Hamburg 5 I am interested only on one way either “Hamburg – Frankfurt” or “Frankfurt – Hamburg” and add them as number of trips made between this two locations. How can i do this in
Numpy multiplication using * (asterisk) returning wrong values when using named variables
I am running into a problem using the operator * with numpy scalars, and it would be great if someone can explain what is going on. Basically, I needed to multiply the sums of columns and rows from various dataframes, and the easiest way to do that was to assign each aggregate to a variable, and then multiply those variables
the speed of numpy sum of different axis
We know that numpy is C order stored so .sum(axis=1) should be faster that .sum(axis=0). But I find that But when the size change to 10000 Answer First of all, I cannot reproduce the effect with the last version of Numpy (dev) nor the version 1.21.5. I got respectively 30.5 ms and 36.5 ms (so the opposite behaviour). Thus, the
storing result from function directly into DataFrame with return
I’m new to programming and python, I’m trying to create a function to iterate over a dataframe and directly store results from the function to dataframe, so far here is what I’ve done: after running it I’m able to get the NumPy array from p and store it to a variable then transform it into dataframe, but that’s only work