Glossary of terms¶
This page lists alphabetically the meaning of various terms used in dislib’s documentation:
term: | environment |
---|
- array-like
- an instance of an object that can be interpreted as an array (e.g., a list, a sparse matrix, a NumPy array, etc.)
- block
- a part of a ds-array that is normally stored remotely.
- column
- a column in a ds-array.
- column block
- a block or set of blocks representing a set of columns in a ds-array.
- csr_matrix
- an instance of a sparse matrix in compressed sparse row format.
- ds-array
- an instance of a distributed array.
- estimator
- anything that learns from data. Typically, an object that fits a model given some parameters and input data.
- feature
- each of the dimensions of a sample array. For example, petal length or color.
- fit
- learn a model from input data.
- label
- a number that represents the category of a sample.
- ndarray
- an instance of a NumPy array.
- predict
- infer the category of unlabeled data according to a fitted model.
- row
- a row in a ds-array (also a sample).
- row block
- a block or set of blocks that represent a set of rows in a ds-array.
- sample
- an array that normally represents an observation or an instance. For example, the characteristics of a particular flower.
- shape
- the total number of rows and columns of a ds-array (or NumPy array)
- spmatrix
- an instance of a SciPy’s sparse matrix.
- synchronization
- when the execution of a parallel application stalls until certain data is generated, or until all tasks have finished.
- task
- a unit of computation that can be executed in a remote computer.