HTTPS clone URL
Subversion checkout URL
Open source framework for predictive modeling on Apache Hadoop
Java GAP Other
Fetching latest commit...
Cannot retrieve the latest commit at this time.
|Failed to load latest commit information.|
About Nectar Nectar is a collection of building blocks to create high-performance and massively scalable business analytics solutions working on big data. Business analytics solutions, be it customer churn prediction, behavior analysis, fraud prevention etc, rely on techniques from statistics, mathematics, machine learning etc to discover and present insights and recommend actions based on complex, sometimes not very intuitive, analysis. The complexity in developing such solutions typically arises from two factors: Solution development ·Discerning / applying existing algorithms or discovering new algorithms for data treatment and predictive analysis ·Implementing the algorithms in a high-level language. Big Data handling ·Handling peta-byte level data in an efficient way ·Distributing processing load in multiple computing nodes Currently, developers, statisticians, system integrators and students face the challenge from both aspects which eventually involves significant learning, development, testing and sometimes dependence on people with specialized skills. In the worst case, these issues drive people / organization away from taking up such projects although they are convinced about strategic values of these. With the availability of Nectar, you have your way now! Nectar assists you to get started and making enterprise-grade business analytics solutions in less than 3 days. It just needs basic familiarity with Hadoop and Map-Reduce paradigm. Knowledge of statistical methods and optimization techniques are not assumed. Now you have the power to combine building blocks the way you like and create innovative solutions or implement your own algorithm in a big data environment. Nectar is open source and distributed with Apache license. Nectar primitives are efficient and manage job creation and dependency management within themselves. A flat delimited file is expected as input. In addition to Java based APIs, the distribution release contains a shell like facility where Nectar primitives can be executed interactively. Using Nectar 1. Read INSTALL.txt to know installation steps 2. Please go through the documents in youshouldreadthis folder. 3. Refer to javadocs in doc folder for Nectar APIs. Happy Analytic Journey with Nectar !!