1. Sparse matrix creation: coo_matrix()
from import coo_matrix # Build sparse matrices data = [1,2,3,4] row = [3,6,8,2] col = [0,7,4,9] c = coo_matrix((data,(row,col)),shape=(10,10)) #Construct a 10*10 sparse matrix where the values and positions that are not 0 are in the first parameter print(c)
2. Sparse matrices into dense matrices: todense()
d = () print(d)
3. Convert a matrix with many 0-values into a sparse matrix
e = coo_matrix(d) #Convert a matrix with many values of 0 to a sparse matrix print(e)
4. save: similar to the .mat format in matlab, python can also save the parameter data, in addition to save as csv, json, excel, etc., I personally think that matlab's .mat format is really strong, what can be saved directly ~~~
import numpy as np # (arg_1,arg_2)，arg_1is the filename，arg_2is the array to be saved
aa = (d)
print(aa)
# save
('test_save_1.npy', aa) # Keep an array
('test_save_2', aa=aa, d=d) #Saving multiple arrays，where sparse matrices can be saved directly
5. load: load parameter data
#load
a_ = ('test_save_1.npy')
print(a_)
dt = ('test_save_2.npz') The #npz data is loaded as a dictionary format data
print(dt)
print(dt['aa'])
print(dt['d']) #Get the value of one of the parameters，Similar to the dictionary form of getting
6. Parameter name for getting npz data
# Get the name of the parameter p_name =list(()) print(p_name) # Get the value p_value =list(()) print(p_value)
This is the whole content of this article.