We are pleased to announce the inception of the Intel® Data Analytics Acceleration Library (Intel® DAAL) open source project. Intel DAAL helps to speed up big data analysis by providing highly optimized algorithmic building blocks for all stages of data analytics (preprocessing, transformation, analysis, modeling, validation, and decision making) in batch, online, and distributed processing modes of computation. The open source project is licensed under Apache License 2.0.
Please see an overview video, blog posts, etc.: https://01.org/daal
- C++ and Java APIs.
- Support Linux*, Windows*, and OS X*.
- Batch, online, and distributed processing modes.
- Data connectors to HDFS, MySQL, CSV files, in-memory strings, as well as user defined data sources.
- Descriptive statistics algorithms: low order moments, quantiles, correlation and variance-covariance matrix, cosine distance matrix, correlation distance matrix.
- Data transformation: QR decomposition, Cholesky decomposition, SVD.
- Linear regression
- Classification algorithms: SVM, Naive Bayes, ensemble learning with boosting (AdaBoost, BrownBoost, LogitBoost), multi-class classifiers.
- Unsupervised learning: PCA, Kmeans, expectation-maximization for GMM, outlier detection, association rules mining.
- Recommendation system: collaborative filtering based on ALS (alternating least squares).
- Support for neural networks including convolutional neural networks.
You can find the current Intel DAAL documentation on the Intel(R) Data Analytics Acceleration Library 2017 Beta Documentation Download web page.
Some tutorials and code samples of how to use Intel DAAL can be found on this page.