I want to store points in an array, but I am not sure how to iterate thorugh it.
x = np.dot(weights, corners)
x_points = np.zeros([100, 2])
for i in range(101):
x_points = (x + corners[np.random.randint(3)]) / 2
x = x_points
I have tried to iterate my array like this:
x = np.dot(weights, corners)
x_points = np.zeros([100, 2])
for i in range(101):
x_points[i:,] = (x + corners[np.random.randint(3)]) / 2
x = x_points
How do i store each point in my array?
You for loop is meant to iterate over your number of points, but you never specifically pick the i-th row from the original x array to work on; instead, you just use the entire array on every iteration. Here's what you should be doing:
for i in range(101):
x_points[i,:] = (x[i,:] + ...) / 2
(Note that you could even omit the second index if you wanted to: x[i] is the same as x[i,:].)
In addition, you indexing for storing in the x_points array is wrong. I'm assuming you meant to write x_points[i,:] like I did above, instead of x_points[i:,]. The latter slices the array from the i-th row to the end along the first axis, so your code would actually modify more than just one row of x_points at a time.
Finally, I don't understand why you are assigning x = x_points on every loop iteration. This way, after the loop runs for the first time, your original x array doesn't exist anymore. If you want to "rename" the finalized x_points, you need to place this line after the for loop.
x[i+1] now? Are you trying to replace points below the current row that you're operating on? You never mentioned that. I'm still not entirely sure what you are trying to achieve.corners, but why are you trying to index it again? Don't you just want to add that row to the current one from x?Using Numpy
Assuming you have only 10 points to keep it readable here, I have shown how to create the arrays using Numpy, and accessing its elements.
>>> weights = np.random.randn(10,4)
>>> corners = np.random.rand(4,2)
>>> x=np.dot(weights, corners)
>>> x
array([[ 0.41183813, 0.18830351],
[ 0.10599174, 0.76246711],
[ 0.50235149, 2.76642114],
[ 0.17072047, 0.67126034],
[-0.25400796, -1.20589651],
[-0.04360992, -0.06102023],
[ 0.0446144 , 0.48355716],
[ 0.8501312 , 1.93899732],
[ 0.44656254, 1.05180096],
[ 0.00397146, 0.19248246]])
>>> x_points = np.zeros([10, 2])
Now you have a set of x_points, initialized to 0.0, and your dot product of weights and corners in x.
for i in range(len(x_points)):
x_points[i] = (x[i] + corners[np.random.randint(3)]) / 2
All your randomly initialized points, that took x values, and the corners into consideration have been written to x_points.
>>> x_points
array([[ 0.33206919, 0.25078735],
[ 0.17914599, 0.53786916],
[ 0.46472124, 1.74498146],
[ 0.29890573, 0.69740106],
[-0.08596056, -0.17476289],
[ 0.01923846, 0.39767525],
[ 0.14845732, 0.39841418],
[ 0.46610902, 1.39768402],
[ 0.43682676, 0.88767137],
[ 0.04302915, 0.52442659]])
For simple Python 2D arrays
You create a 2D array as follows:
matrix = [[0 for x in range(num_cols)] for y in range(num_rows)]
Then, you access it like you would access a list with the index:
matrix[2][3] = 23
for i in range(num_rows):
for j in len(num_cols):
print(f"The element in the {i}th row, {j}th col is {matrix[i][j]}")
xis a 100x2 array of points, and you want to average it with some random values fromcorners?