Cloud 1305

Cloud1305 beta* - The Analytics1305 Machine Learning Library on the Amazon Public Cloud Service

The Amazon public cloud service, EC2, allows anyone with an internet connection to rent computing resources at the cost of a small hourly fee. This has bought vast computing resources to the mass market- anyone with an internet connection can now rent multiple machines, from single core's with less than 2GB of RAM to larger boxes with 8 cores and 68 GB of main memory. Before one would have to incur significant expense if requiring more computing muscle necessary these days in many fields. One such is that of machine learning, statistics and data mining which when translated to the software industry comprises the field of advanced analytics.

Analytics1305 has built a sophisticated library of machine learning and statistics methods that have been designed from the ground up to be scalable and extremely fast. It contains many of the latest and most advanced algorithms available for the common data analysis tasks of classification, regression, clustering etc. Born in a lab and then professionally authored in solid C++ code, our framework presents the first commercially available and supported software to be able to guarantee first-class accuracy and measurable speed performance with innovative techniques built in to make the analysis of large volumes of data feasible and more manageable.

Combining our library with the Amazon EC2 service has given us the opportunity to facilitate its public use. Within a few minutes and a free EC2 account later the analyst can launch multiple machines on the cloud with our libraries pre-installed. Then the data can be uploaded and the required analysis task run. We will provide free support through our documentation, blogs and forums.

List of Algorithms
  1. All Neighbors beta*
    1. Nearest/Furthest
    2. K-neighbors/Range-neighbors
    3. Exact/Approximate
    4. Batch/Progressive
    5. Single/Dual tree
    6. Euclidean/Weighted Euclidean
    7. Sparse/Dense/Categorical
  2. Kernel Density Estimation beta*
    1. Gaussian/Epanechnikof kernel
    2. Single/Dual tree
    3. Exact/Approximate
    4. Batch/Progressive
    5. Euclidean/Weighted Euclidean
    6. Sparse/Dense/Categorical
  3. K-means beta*
    1. Tree/Naive
    2. Euclidean/Weighted Euclidean
    3. Sparse/Dense/Categorical
  4. Support Vector Machines beta*
    1. Sparse/Dense/Categorical
    2. Gaussian/Polynomial kernel
    3. Bootstrap (map-reduce) version
  5. More to come, subscribe to our newsletter to get updates

The service is free and we intend to keep it free for the basic operations running on a single machine . You will only pay Amazon charges for whatever you use directly to Amazon. Please refer to our getting started guide for how to start running Amazon Cloud machines and our documentation, (pdf version).