Softmax Function
Converts a vector of numbers into probabilities that sum to 1.
Notes tagged MachineLearning
-
-
Sigmoid Function
Squeezes any number into a range between 0 and 1.
-
Domain Shift
When production data diverges significantly from the training dataset
-
Out-of-Domain
When the production data is significantly different from the training distribution.
-
-
-
Mean Absolute Error
The average absolute difference between predictions and labels, less sensitive to outliers than RMSE.
-
Root Mean-Squared Error
A regression metric that penalises large errors.
-
Hyperplane
a subspace where the dimension is one less than the ambient space
-
Principal Component Analysis
a technique for reducing the dimensionality of data
