Intel® Data Analytics Acceleration Library (Intel® DAAL) is the library of Intel® architecture optimized building blocks covering all stages of data analytics: data acquisition from a data source, preprocessing, transformation, data mining, modeling, validation, and decision making.
If you are new to Intel DAAL, start with the summary of the product functionality.
Algorithms implemented in the library support batch, online, and distributed processing modes of computations. More information.
Algorithms implemented in Intel DAAL include:
- Algorithms for analysis:
- Association rules
- Correlation and variance-covariance matrices
- Distance matrices:
- Density-Based Spatial Clustering of Applications with Noise
- K-Means clustering
- Math functions:
- Matrix decompositions:
- Moments of low order and quantiles
- Iterative solvers:
- Objective functions:
- Normalization algorithms:
- Outlier detection:
- Principal component analysis and Principal Component Analysis Transform.
- Quality metrics for classification algorithms and linear regression
- Sorting observations by features
- Algorithms for training and prediction:
- Classification and regression:
- Implicit alternating least squares recommendation system
Starting from Intel® DAAL 2020, The Neural Networks will not have any new features and functionalities.
The support will be completely discontinued from Intel® DAAL 2021.
Intel recommends developers switch to and use other Intel® Optimized Deep Learning frameworks (TensorFlow*, PyTorch*, PaddlePaddle*, MXNet*, BigDL and Caffe*) and Intel® AI Software libraries (NGRAPH™, Intel® MKL, Intel® MKL-DNN, Intel® CLDNN, etc.) for Neural Networks.
Intel® DAAL provides application programming interfaces for C++, Java*, and Python* languages. Visit Intel® Data Analytics Acceleration Library API Reference to download API References for C++, Java*, and Python*. API Reference for C++ is also available online on IDZ, see C++ API Reference for Intel® Data Analytics Acceleration Library.