Compares two matrices for approximate equality.
Compares two matrices for approximate equality.
A matrix.
A second matrix.
equality threshold
True iff all corresponding elements of the two matrices are within thresh
of each other.
Compares two arrays for approximate equality.
Compares two arrays for approximate equality. Modify margin by setting Stats.thresh.
A array of numbers.
A second array of numbers.
equality threshold
True if the two numbers are within thresh
of each other.
Compares two numbers for approximate equality
Compares two numbers for approximate equality
A number.
A second number.
equality threshold
True if the two numbers are within thresh
of each other.
Computes top-k classification error based on a distribution of predictions.
Computes top-k classification error based on a distribution of predictions.
Distribution of predictions.
Distribution of actuals.
How many classes to include.
The error percent.
Computes error percent based on array of predicted labels and array of actual labels.
Computes error percent based on array of predicted labels and array of actual labels.
The predicted labels.
The actual labels.
Total number of examples.
Error percent.
Given a local matrix, compute the mean and variance of each row.
Given a local matrix, compute the mean and variance of each row. Subtract the row mean from each row and divide by the sqrt(variance + alpha).
Input matrix.
Alpha for the denominator.
Normalized Matrix.
Margin to use for comparing numerical values.