OperationsResearch.com

Archive for the ‘Predicitive Models’ Category

Predictive analytics is a solution used by many businesses today to gain more value out of the large amounts of raw data by applying techniques that are used to predict future behaviors within an organization, it’s customer base, it’s products and services. Predictive analytics encompasses a variety of techniques from data mining, stastics and game theory that analyze current and historical facts to make predictions about future events.

The benefits of implementing predictive analytics is undeniable. There are countless documented case studies and success stories where predictive analysis yielded a substantial return on investment, helped companies optimize existing processes, provided a better understanding of customer behavior, identified unexpected opportunities, and anticipated problems before they occurred. But with all of the benefits associated with predictive analytics, there are many challenges that accompany becoming an analytics-driven organization.

The perceived complexity is the largest challenge facing executives today. The cost of implementation is a close second. While these are legitimate fears, many tools are being developed to simplify the process and establish transparency from the complex formulas and statical modeling. It is, however, up to organizations to educate themselves on the basics and concepts of predictive analysis in order to fully utilize these tools.

Another challenge, which is more technical, is the traditional approach of having analyst explore data sets by saving data and manually applying relationships in order to make predictive assumptions. While this can work at a basic level of predictive analytics, predictive analytics at it’s most effective application requires extremely large amounts of data and thus is best suited for analytics platforms wih parallel processing, which support custom analytical applications that query data using SQL.

This brings us to another challenge with implementing predictive analytics in your organization, and that is managing the enormous data volumes associated with it. Some organizations known to apply leading edge analytical techniques, are gathering perabytes (that’s approximately 1000 terabytes, or 1 million gigabytes) of data. While these amounts of data require costly data warehouse upgrades, it enables organizations to form very comprehensive analytics and it enhances visitor/customer experience by providing targeted, customized marketing and services.

But with these large amounts of data and data storage comes the challenges of producing the platform for processing this data with complex formulas at fast rates. Because of this, analytic platforms often run off massively parallel processing (MPP) databases. MPP databases coordinate processing of a single program by more than one processor by dividing up parts of a program into several processors with their separate memory and operating systems. But many organizations that cannot afford MPP databases, instead implement analytical platforms as data marts to off-load complex processing.

While these challenges to indeed appear to be complex, the important thing to know is that if you have the architecture to support it, there are several tools out there that take out the complexities and applying predictive modeling.

Product Description
This digital document is an article from Indiana Business Magazine, published by Thomson Gale on October 1, 2005. The length of the article is 552 words. The page length shown above is based on a typical 300-word page. The article is delivered in HTML format and is available in your Amazon.com Digital Locker immediately after purchase. You can view it with any web browser.

Citation Details
Title: Crystal ball of health car… More >>

Crystal ball of health care? Can predictive modeling help lower costs?: An article from: Indiana Business Magazine

In this session, Jennifer Thompson, MS, introduces data cleaning and outliers. Outliers can be a tricky problem for a data mining project. This session will address these problems and help understand what caused them in the first place.

Welcome to StatSoft’s Introduction to Data Mining video series. The series covers hands on tutorials of data mining applications. Subscribe to this series at www.statsoft.com/dmsubscribe to be notified as each future episode becomes available and for any supplementary materials provided for a lesson. … “Data Mining” “predictive modeling” “predictive analytics” “data analysis. analytics” “business intelligence” STATISTICA

Simulations Plus Sees New Market Ahead
Simulations Plus, Inc. announced it will license software that can help scientists predict how drug molecules will react in the human body, a move the company expects will ultimately create an entirely new market for its scientific-modeling software.

Read more on San Fernando Valley Business Journal

How can analytical insights from the sales data help maximize my revenue? How can I maximize the value of my offerings to my loyal customers? Which product sells more and what is the inventory? These are some of the questions that are typically faced by retailers. Retailers also face problems like expensive store locations, slim margins and tenuous customer loyalty. A retailer even needs to consider every category while making smarter decisions and manage their businesses more effectively.

Reporting without analysis simply provides answers to yesterday’s questions. Analytical reporting application merges the benefits of static business reporting with the interactive nature of analysis into one application to get insightful analysis reported throughout the business. Analytical reporting delivers a complete reporting solution surrounded by a very powerful analytic tool coupled with an easy to use environment for today’s business user.

Analytical tools and statistical methods help retailers understand the pattern of the data and make business solutions. Let us discuss about analytical services starting with predictive analytics.

Predictive analytics provides the marketer with something beyond standard business reports and sales forecasts i.e., actionable predictions for each unique scenario. These predictions could encompass all channels, both online and off, foreseeing which customers will buy, click, respond, and convert.

Predictive analytical models exploit patterns found in historical transactional data to identify risks and opportunities. Models capture relationships among many factors to allow assessment of risk or potential associated with a particular set of decisions.

Predictive analytics is a data mining technology that uses customer data to build a predictive model specialized for business operations. This analysis requires collective information on customer’s purchase, behaviour and demographics. Predictive models analyze past performance of a customer to improve market effectiveness to assess how likely customer behaves. This process learns from your organization’s collective experiences by leveraging your existing logs of customer purchases, behaviour and demographics.

Retailers can incorporate predictive analytics into their daily operations to improve their business processes. This would enhance a retailer’s decision making and they would gain the ability to direct, optimize, and automate decisions, to meet defined business goals.  By using predictive analytics, they can increase their probability and decision making processes.

Predictive analytics provides a quantitative foundation for rapidly identifying, objectively evaluating and confidently pursuing new market opportunities. It helps retail organizations not only to grow and generate revenue quickly and predictably but also improve the operating performance and enable retailers to make more accurate forecasts, manage resources and improve sales productivity.

Seetarama Hegde
Retail Consultant

CustoLogix Solutions

URL | www.custologix.com

CustoLogix with its wide experience in statistical analysis helps retailer to improve retail profitability through Analytics. To know more about predictive analytics please visit CustoLogix at www.custologix.com/service

Product Description
This is the collected papers presented at the Integrating Safety and Environmental Knowledge Into Food Studies towards European Sustainable Development (ISEKI) workshop on risk assessment in the food industry…. More >>

Predictive Modeling and Risk Assessment

Predictive analytics is a solution used by many businesses today to gain more value out of large amounts of raw data by applying techniques that are used to predict future behaviors within an organization, it’s customer base, it’s products and services. Predictive analytics encompasses a variety of techniques from data mining, stastics and game theory that analyze current and historical facts to make predictions about future events.

Predictive models examine patterns found in historical and transactional data to identify opportunities and risks. Predictive models capture relationships among many factors to allow assessment of risk or potential associated with a particular set of conditions, guiding decision making for candidate transactions.

There are some basic and more complex predictive analytics techniques. Three basic techniques include:

-Data Profiling and Transformations
-Sequential Pattern Analysis
-Time Series Tracking

Data profiling and transformations are functions that analyze row and column attributes and dependencies, change data formats, merge fields, aggregate records, and join rows and columns.

Sequential pattern analysis discovers relationships between rows of data. Sequential pattern analysis is used to identify frequently observed sequential occurrence of items across ordered transactions over time. Such a frequently observed sequential occurrence of items (called a sequential pattern) must satisfy a user-specified minimum support. Understanding long-term customer purchase behavior is an example of the sequential pattern analysis. Other examples include customer shopping sequences, click-stream sessions, and telephone calling patterns.

Time series tracking tracks metrics that represent key behaviors or business strategies. It is an ordered sequence of values of a variable at equally spaced time intervals. Time series analysis accounts for the fact that data points taken over time may have an internal structure (such as autocorrelation, trend or seasonal variation) that should be accounted for. Examples include patterning customer sales that indicate product satisfaction and buying habits, budgetary analysis, stock market analysis, census analysis, and workforce projections.

More advanced predictive analytics techniques include:

-Time Series Forecasting
-Data Profiling and Transformations
-Bayesian Analytics
-Regression
-Classification
-Dependency or Association Analysis
-Simulation
-Optimization

Time series forecasting predicts the future value of a measure based on past values. Time series forecasting uses a model to forecast future events based on known past events. Examples include stock prices and sales revenue.

Data profiling and transformation uses functions that analyze row and column attributes and dependencies, change data formats, merge fields, aggregate records, and join rows and columns.

Bayesian analytics capture the concepts used in probability forecasting. It is a statistical procedure which estimate parameters of an underlying distribution based on the observed distribution. An example is used in a court setting by an individual juror to coherently accumulate the evidence for and against the guilt of the defendant, and to see whether, in totality, it meets their threshold for ‘beyond a reasonable doubt’.

Regression analysis is a statistical tool for the investigation of relationships between variables. Usually, the investigator seeks to ascertain the causal effect of one variable upon another-the effect of a price increase upon demand, for example, or the effect of changes in the money supply upon the inflation rate.

Classification used attributes in data to assign an object to a predefined class or predict the value of a numeric variable of interest. Examples include credit risk analysis, likelihood to purchase. Examples include acquisition, cross-sell, attrition, credit scoring and collections.

Clustering or segmentation separates data into homogeneous subgroups based on attributes. Clustering assigns a set of observations into subsets (clusters) so that observations in the same cluster are similar. An example is customer demographic segmentation.

Dependency or association analysis describes significant associations between data items. An example is market basket analysis. Market basket analysis is a modeling technique based upon the theory that if you buy a certain group of items, you are more (or less) likely to buy another group of items.

Simulation models a system structure to estimate the impact of management decisions or changes. Simulation model behavior will change in each simulation according to the set of initial parameters assumed for the environment. Examples include inventory reorder policies, currency hedging, military training.

Optimization models a system structure in terms of constraints to find the best possible solution. Optimization models form part of a larger system which people use to help them make decisions. The user is able to influence the solutions which the model produces and reviews them before making a final decision as to what to do. Examples include scheduling of shift workers, routing of train cargo, and pricing airline seats.