XLMiner® SDK Platform

Capabilities

• Forecasting, data mining, text mining in Excel
• Uncover new revenue, predict churn or fraud
• Explore, partition, transform data, extract features
• Use regression, trees, neural nets, ensembles, PMML
• Use with C++, C#, Java, Python or R

XLMINER SDK

XLMiner SDK offers developers working in C++, C#, Java, Python or R a powerful, high-level API for predictive analytics.

•  Data Exploration and Feature Selection
• Unsupervised Machine Learning
• Classification and Prediction Algorithms
• Time Series Forecasting Methods
• DataFrame, Estimator, Model Objects
•  Pipeline Multiple Operations
• Run on 32-bit or 64-bit Windows
• Exploit multiple cores automatically
• Run on desktop, server, cloud
• Connect to Spark Big Data clusters

FEATURES

Everything You Need to Develop and Deploy Powerful Predictive Analytics Applications Common, Easy-to-Use Object-Oriented API Use the same high-level objects (like DataFrame, Estimator and Model), properties and methods across different programming languages. Parallelized Data Mining Algorithms Exploit multiple processor cores using XLMiner SDK's built-in parallelized algorithms, without having to write parallel code yourself. Thread-safe for Multi-user Applications Call XLMiner SDK on multiple concurrent threads -- making it easy to build Web server and other applications that handle multiple clients concurrently. Example Applications in Multiple Languages Source code is included in each supported programming language for example applications, illustrating how to use the full range of XLMiner SDK features. Help and Support Our xxx-page XLMiner SDK User Guide and our highly-regarded technical support supports your efforts. Straightforward Runtime Licensing Obtain licenses to distribute XLMiner SDK with your application, or use it on your on-premise or cloud-based server(s), with simple terms that work for your application.

ALGORITHMS

Fast Text Processing Use stemming, term normalization, and vocabulary reduction to create a term-document matrix with weightings. Latent Semantic Indexing Extract concepts from text, representing common clusters of words in documents to create a concept-document matrix. Unsupervised machine learning Use k-Means Clustering and Hierarchical Clustering to structure rows, and Principal Components to reduce columns. Powerful Feature Selection Use filter, wrapper and embedded methods to identify variables with the greatest predictive power. Exponential Smoothing and ARIMA models Use exponential smoothing and ARIMA (Auto-Regressive Integrated Moving-Average) methods, with or without seasonality. Linear and Logistic Regression Use multiple linear regression and logistic regression, with automatic handling of categorical variables, automatic variable screening, and forward, backward, stepwise and exhaustive variable selection methods. Ensembles of "Weak Learners" Use boosting, bagging, and "random forest" methods to combine results from many instances of different algorithms. Classification and Regression Trees Use regression-based "decision trees" for both classification and prediction tasks. Neural Networks Use multi-layer feedforward neural networks for both classification and prediction tasks. Discriminant Analysis and Naive Bayes Use these classic, but surprisingly powerful methods for classification. Association Rules Automatically construct sets of rules for "market basket" and recommendation systems.