1. Array Creation and Initialization
np.array()
Creates a NumPy array from lists or tuples.
np.zeros(shape)/np.ones(shape)
Creates arrays filled with zeros or ones, often used for initializing weights or placeholders.
np.empty(shape)
Creates an array without initializing values (useful for pre-allocated storage).
np.full(shape, value)
Creates an array filled with a specified constant value.
np.arange(start, stop, step)
Generates an array of evenly spaced values within a given interval.
np.linspace(start, stop, num)
Creates an array of
num evenly spaced values between start and stop.np.eye(n)/np.identity(n)
Creates an identity matrix, commonly used in linear algebra.
2. Reshaping and Manipulating Arrays
array.reshape(shape)
Reshapes an array without changing its data.
array.flatten()
Flattens a multi-dimensional array to a one-dimensional array.
np.transpose(array)orarray.T
Transposes an array, swapping rows and columns.
np.concatenate((a, b), axis=0/1)
Concatenates two arrays along a specified axis.
np.vstack()/np.hstack()
Stacks arrays vertically or horizontally.
np.split(array, sections)
Splits an array into multiple subarrays, useful for dividing data into batches.
3. Indexing and Slicing
- Basic Indexing
Access specific elements using
array[i] or array[i, j].- Slicing
Use
array[start:stop:step] to extract subarrays.- Boolean Indexing
Select elements based on a condition:
array[array > 0]
- Fancy Indexing
Select specific elements using an array of indices:
array[[1, 2, 4]]
4. Mathematical and Statistical Operations
- Element-wise Operations
Use standard operators (
+, -, *, /) for element-wise calculations.- Aggregations
np.sum(array, axis=None): Computes the sum across specified axes.np.mean(array, axis=None): Calculates the mean across specified axes.np.std(array, axis=None)/np.var(array, axis=None): Standard deviation and variance.
- Min/Max Operations
np.min()/np.max(): Finds minimum and maximum values.np.argmin()/np.argmax(): Finds indices of min/max values.
- Element-wise Functions
np.exp(array): Exponential function.np.log(array): Natural logarithm.np.sqrt(array): Square root.
- Matrix Operations
np.dot(a, b): Dot product of two arrays.np.matmul(a, b)ora @ b: Matrix multiplication.
5. Random Number Generation
np.random.rand(d0, d1, ...)
Generates random values between 0 and 1 with the given shape.
np.random.randint(low, high, size)
Generates random integers between
low and high.np.random.randn(d0, d1, ...)
Generates random numbers from a standard normal distribution.
np.random.seed(seed)
Sets the random seed for reproducibility.
6. Linear Algebra and Matrix Operations
np.linalg.inv(array)
Computes the inverse of a matrix.
np.linalg.det(array)
Computes the determinant of a matrix.
np.linalg.eig(array)
Computes the eigenvalues and eigenvectors of a matrix.
np.linalg.svd(array)
Performs singular value decomposition (useful for dimensionality reduction).
np.linalg.norm(array)
Computes the norm (magnitude) of an array, useful for vector normalization.
7. Broadcasting
- What is Broadcasting?
Enables NumPy to perform operations on arrays of different shapes by “stretching” them to compatible shapes.
- Example
Adding a 1D array to a 2D array row-wise:
array_2d + array_1d
8. Sorting and Searching
- Sorting
np.sort(array): Sorts an array in ascending order.np.argsort(array): Returns the indices that would sort the array.
- Conditional Selection
np.where(condition, x, y): Chooses elements fromxorybased on a condition.
- Unique Elements
np.unique(array): Finds unique elements in an array.
9. Statistical Sampling
np.random.choice(array, size, replace)
Samples elements from an array with or without replacement.