SPSS for BigData Modeling

1 Featured IBM Analytics products

SPSS\|Predictive analytics	IBM SPSS predictive analytics software offers advanced techniques in an easy-to-use package to help you find new opportunities, improve efficiency and minimize risk.
IBM DB2\|Database	IBM DB2 is the database of choice for enterprise-wide solutions. Optimized to deliver industry-leading performance while lowering costs, IBM DB2 offers extreme performance, flexibility, scalability and reliability for any size organization.
Cognos Analytics on Cloud\|Business Intelligence	Get the self-service you expect, data governance you require, and reporting you trust with a secure business intelligence software-as-a-service (SaaS) solution.
Hadoop\|Data management platform	IBM® BigInsights™ for Apache™ Hadoop is an industry standard Hadoop offering that combines the best of open source software with enterprise-grade capabilities.

2 SPSS Analytic for BigData

2.1 分析服务器3.0

结合使用 IBM SPSS Modeler 与 IBM SPSS Analytic Server 时，分析人员可以在大数据上开发并部署预测性分析，而无需昂贵的技术技能。IBM SPSS Analytic Server 可提供：

使用大数据系统的以数据为中心的开放式集成架构
支持各种流行的 Hadoop 版本，例如，IBM InfoSphere® BigInsights™、Cloudera、Hortonworks 和 Apache Hadoop
已定义的接口，以合并旨在用于数据的新统计算法
熟悉的 IBM SPSS 用户界面，隐藏大数据环境的详细信息，以便分析人员可以专注于数据分析
可通过扩展来解决几乎任何规模问题的解决方案

SPSS Analytic Server Architecture

2.1.1 Analytic Server 管控台

Analytic Server provides a thin client interface for managing data sources and projects.

A data source is a collection of records, plus a data model, that define a data set for analysis. The source of records can be a file (delimited text, fixed width text, Excel) on HDFS, a relational database, HCatalog, or geospatial.

Projects are workspaces for storing inputs and accessing outputs of jobs. They provide the top-level organizational structure for containing files and folders. Projects can be shared with individual users and groups.

2.1.2 SPSS Modeler 集成

SPSS® Modeler is a data mining workbench that has a visual approach to analysis. Each distinct action in a job, from accessing a data source to merging records to writing out a new file or building a model, is represented by a node on the canvas. We link these actions together to form an analytic stream. To build an analytic stream that runs with Analytic Server

受支持的节点

A stream that begins with anything other than an Analytic Server source node will be run locally.

Graphs
All Graph nodes are supported.
Modeling
The following Modeling nodes are supported: Time Series, TCM, Tree-AS, C&R Tree, Quest, CHAID, Linear, Linear-AS, Neural Net, GLE, LSVM, TwoStep-AS, Random Trees, STP, and Association Rules. Further notes on those nodes’ functionality follow.

案例

基于 Hadoop 平台下 Modeler 的大数据应用剖析

2.2 Spark 和分析服务器

分析服务器与 Apache Spark 集成以提高性能。

何时使用和不使用 Spark

如果在 Hadoop 集群中将 Spark 作为 Ambari 服务进行安装，那么分析服务器使用它来处理大数据作业。以下准则适用于确定何时不使用 Spark。

1 如果数据集小于 128 MB，那么分析服务器在分析服务器 JVM 中使用嵌入式 MapReduce 函数并且不使用 Spark 或 Hadoop 集群。
2 如果未在集群上安装 Spark，那么分析服务器使用 MapReduce v2。
3 分析服务器使用 MapReduce v2 来构建 PSM 模型。作业以 PSM 模型构建结束时，分析服务器使用 Spark 通过导致模型构建的所有步骤来处理作业，并写入磁盘，然后使用 MapReduce 来构建 PSM 模型。例如，如果作业包含后面跟着 PSM 模型构建的连接，那么在 Spark 中运行连接并在 MapReduce 中的连接数据上运行 PSM。

案例

SPSS Algorithms Optimized for Apache Spark & Spark Algorithms Extending SPSS Modeler

2.3 使用 SPSS + BigInsights 共同构架大数据分析平台

3 Supported Language with SPSS Modeler 18.0

ModelerUsersGuide.pdf

3.1 R

IBM® SPSS® Modeler supports R.

f you have a compatible copy of R installed, you can connect to it from IBM SPSS Modeler and carry out model building and model scoring using custom R algorithms that can be deployed in IBM SPSS Modeler. You must also have a copy of IBM SPSS Modeler - Essentials for R installed. IBM SPSS Modeler - Essentials for R provides you with tools you need to start developing custom R applications for use with IBM SPSS Modeler.

参考链接