Friday Jun 03, 2016

Using Forecast Function in Oracle DV / OBIEE 12c

Oracle DV and OBIEE 12c offer a right click interaction to include forecast data in many visualizations.

But it also let users manually configure and edit Forecast functions as custom calculations. This Forecast function is based on a R script ran by Oracle BI Server, and comes with various options and parameters enabled. For example, it allows to define the number of periods to forecast, the type of forecasting model to use (Arima, ETS) and what specific parameters for this model (error type, seasonality, trending, Box Cox, Damping...) as well as what output data to produce (value, confidence levels high and low bounds). All these options combine in a range of various possibilities for calculating the most appropriate forecast information for the use case.

A free pre-built DV Destkop project showing various combinations of these options has been posted on the Oracle Technology Network Dataviz example page here (scroll to project example 'Forecast Syntax Examples') and this YT video gives a very brief introduction of how to use this function :

The FORECAST() Logical SQL Function takes a measure parameter e.g. revenue and a variable list of time dimensions. Optional column aliases can also be used. The default FORECAST XML (filerepo://obiee.TimeSeriesForecast.xml) script file on your server or laptop can be overridden by specifying a new one in the options string. 

Syntax : FORECAST( <measure_expr>, (<time_dimension_expr>), <column_name> , <options>, [<runtime_binded_options>])

- measure_expr represents the measure, e.g. revenue data to forecast.
- time_dimension_expr the time dimension to forecast. One or more columns may be provided.
- column_name represents the output column name for forecast.
- options is a string list of name/value pairs separated by ';'

Option Name



numPeriods the number of periods to forecast Integer
predictionInterval

the confidence for the prediction

Integer (1 to 99)

modelType

the model to use for forecasting

ARIMA, ETS

useBoxCox

if TRUE use box cox transformation

TRUE, FALSE

lambdaValue

the Box-Cox transformation parameter. Ignored if NULL or FALSE.

TRUE, FALSE
trendDamp

(ETS model). if TRUE, use damped trend, ie reduce effect of recent trends.

TRUE, FALSE
errorType

(ETS models) : controls how the nearest prior periods are weighted in the output

additive('A'), multiplicative('M'), automatic('Z')

trendType

(ETS models) : controls how the effect of trend is modeled in the output

None('N'), additive('A'), multiplicative('M'), automatic('Z')

seasonType

(ETS models) : controls how seasonal effects are affecting the model outputs.

None('N'), additive('A'), multiplicative('M'), automatic('Z')

modelParamIC Information criterion to be used in comparing and selecting different models and select the best model. 'ic_auto', 'ic_aicc', (corrected Akaike IC),'ic_bic‘ (Bayesian IC), 'ic_auto'(default)

- runtime_binded_options is an optional comma separated list of runtime binded colums and options.

Some Examples from the DVD Project :

Selecting prediction confidence interval, and showing low end and high end bounds

ETS vs ARIMA : what are the differences ?

Playing with ETS Trending and Seasonality parameters, what does it mean :

Find out more by downloading the example here (scroll to project example 'Forecast Syntax Examples').

Thank you.

Wednesday Feb 26, 2014

Oracle repeats as BI and Analytics Leader in Gartner MQ 2014

For the 8th consecutive year, Oracle is a Leader in Gartner’s Magic Quadrant for Business Intelligence and Analytics Platform. Gartner declares that “the BI and analytics platform market is in the middle of an accelerated transformation from Business Intelligence (BI) systems used primarily for measurement and reporting to those that also support analysis, prediction, forecasting and optimization.” Oracle offers all these wide-ranging capabilities across Business Intelligence Foundation Suite, Advanced Analytics and Real-Time Decisions.

Gartner specifically recognizes Oracle as a Leader for several key reasons. Oracle customers reported among the largest BI deployments in terms of users and data sizes. In fact, 69% of Oracle customers stated that Oracle BI is their enterprise BI standard. The broad product suite works with many heterogeneous data sources for large-scale, multi-business-unit and multi-geography deployments. The BI integration with Oracle Applications, and technology, and with Oracle Hyperion EPM simplifies deployment and administration. Not cited in the Gartner report is that Oracle BI can access and query Hadoop via a Hive Oracle Database Connector eliminating the need to write MapReduce programs for more efficient big data analysis.

“The race is on to fill the gap in governed data discovery,” professes Gartner. In this year’s MQ, all the Leaders have been moved “westward,” to the left, to open up white space in the future for vendors who address “governed data discovery” platforms that address both business users’ requirements for ease of use and enterprises’ IT-driven requirements, like security, data quality, and scalability. Although in Gartner’s view no single vendor provides governed data discovery today, Oracle Endeca Information Discovery 3.1, which became available in November 2013 after Gartner conducted the MQ report, is a complete enterprise data discovery platform that combines information of any type, from any source, empowering business user independence in balance with IT governance. Users can mash-up personal data along with IT-provisioned data into easy to use visualizations to explore what matters most to them. IT can manage the platform to meet data quality, scalability and security requirements. Users can benefit from additional subject areas and metadata provided by integration with Oracle BI.

Gartner additionally cites other Oracle strengths such as more than 80 packaged BI Analytic Applications that include pre-built data models, ETL scripts, reports, and dashboards, along with best practice, cross-functional analytics that span dozens of business roles and industries. Lastly, Oracle’s large, global network of BI application partners, implementation consultants, and customer install base provide a collaborative environment to grow and innovate with BI and analytics. Gartner also cites the large uptake in Oracle BI Mobile enabling business users to develop and deliver content on the go.

Tuesday Apr 09, 2013

Big Data Analytics - Advanced Analytics in Oracle Database

That's the title of a new white paper we've just posted. From the executive summary:

Big data doesn’t only bring new data types and storage mechanisms, but new types of analysis as well. In the following pages we discuss the various ways to analyze big data to find patterns and relationships, make informed predictions, deliver actionable intelligence, and gain business insight from this steady influx of information. 

You can check it out here.

Thursday Jan 31, 2013

Using In-Database Analytics to Predict Fraud

Your data warehouse stores critical data telling you what is happening in your business and sometimes why it’s happening. But you can go beyond understanding why something went wrong. You can use past data to predict the future, correcting problems before they happen. In a recent survey that Oracle did of over 300 C level executives, 93% of them thought that their companies were losing an average of 14% of their total revenue because they couldn’t fully leverage the information they had already collected. One key way to do this (and you’ll hear more about this in a future survey) is to use predictive analytics. Let’s take a quick look at why and how.

Turkcell is a leading mobile phone provider in Turkey, with over 34 million subscribers. And like most mobile providers a majority of those subscribers use pre-paid accounts and pre-paid cards. Money launderers take advantage of this, and losses for this business are of the order of $5 for every $10,000. This may not seem like much, but with billions of transactions, this adds up to millions of dollars a year.

Like other companies, Turkcell examine huge quantities of data and build models that help it identify and ultimately predict and prevent fraudulent transactions.  Unlike many other companies, Turkcell does this analysis in its data warehouse. With 100 TB of compressed data – representing over a petabyte uncompressed – it would take a long time to move that data out of the warehouse and keep it up to date as new data arrived. And the window to stop the next fraudulent transaction might have already closed.

Oracle Advanced Analytics enables you to perform sophisticated predictive analytics inside a data warehouse. You can mine your data directly while it is inside the Oracle Database using either SQL or R language APIs or the Oracle Data Miner SQL Developer “work flow” GUI extension, depending on your need and existing skills. You build models for past behavior and use that to predict future behavior, improving your accuracy with time. And best of all, there’s no need to move the data around which takes time you might not have and also leaves you exposed to security risks. As Turkcell said “...we can analyze large volumes of customer data and call-data records easier and faster than with any other tool”

About

We're taking the pulse of the Business Intelligence and Analytics market based on insights and experiences with colleagues, customers,and partners.

Search

Archives
« June 2016
SunMonTueWedThuFriSat
   
1
2
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
30
  
       
Today