Calculate the mode of a numpy array for n elements

Question

I want to calculate the mode of a big numpy array for every 16 elements in the array. I saved the array in a csv file. I opened the file and for every 16 elements, I want to calculate the mode and save the result in another file. If there are no more than 16 elements, I want to perform majority voting anyway. I uploaded the file of predictions. I tried this code below

Thank you

import csv
from scipy.stats import mode
import numpy as np

with open("prediction-pairs-overall.csv") as csvfile:
    readCSV = csv.reader(csvfile, delimiter=',')

    for i in readCSV:
        values,counts = np.unique(readCSV, return_counts=True)   
        index = np.argmax(counts)  ##maximum values



        mode=values[index]
        print(mode)
        #input()

What is your question? Is there a particular problem with the code you've shown? — mkrieger1, Jun 07 '22 at 08:34
I want to calculate the mode for every 16 elements of a big array and save it in an another file (as a majority voting). I tried the above code but it does not work, there are some errors in the code — Nischay, Jun 07 '22 at 08:37
What are the errors? Why don't you remove them so that your code will work? — mkrieger1, Jun 07 '22 at 08:38
If I print the mode, I cant see the values of mode for the first 16 elements — Nischay, Jun 07 '22 at 08:39
My predictions are splitted in 16 blocks for an image. I want to merge all the 16 blocks and predict for a whole image using majority voting (mode) — Nischay, Jun 07 '22 at 08:41
See https://stackoverflow.com/questions/25614749/how-to-import-csv-file-as-numpy-array-in-python — mkrieger1, Jun 07 '22 at 08:47
You are importing `mode` from scipy.stats but then you are declaring `mode` variable, making it useless. Did you read [mode manual](https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.mode.html)? — Zaero Divide, Jun 07 '22 at 11:45

score 0 · Answer 1 · answered Jun 07 '22 at 12:13

You should calculate unique(i) instead of unique(readCSV).

import csv
import numpy as np

with open("prediction-pairs-overall.csv") as csvfile:
    readCSV = csv.reader(csvfile, delimiter=',')
    for i in readCSV:
        values, counts = np.unique(i, return_counts=True)   
        index = np.argmax(counts)  ##maximum values

        mode = values[index]
        print(mode)

Calculate the mode of a numpy array for n elements

1 Answers1