User Behavior Recovery via Hidden Markov Models Analysis

Size: px
Start display at page:

Download "User Behavior Recovery via Hidden Markov Models Analysis"

Transcription

1 User Behavior Recovery via Hidden Markov Models Analysis Alina Maor, Doron Shaked Hewlett Packard Labs HPE Keyword(s): Hidden Markov Model; predictive analytics; classification; statistical methods; clickstream analysis Abstract: In this report, we propose a method for user-behavior profiling and user-intention prediction based on Hidden Markov Models. User behavior is described by a sequence of user-actions. Note that there is no strict correspondence assumption between action order and user purpose or business process. On the contrary, we assume actions characterizing a purpose or a business process will have order variations or be interleaved with actions emerging from different purposes. The statistical nature of user behavior and the partial order assumption fit the statistical Hidden Markov Model. Different models are trained for different user profiles or to predict predetermined behavior types. Prediction can be incorporated with other predictors in a maximum likelihood setting. The proposed technique is generic and can be applied to profiling of any given number of (predefined) profiles. The only assumption is that profiles are characterized by different dynamics. The method was successfully applied in a twogroup scenario to actions on HP.COM for predicting eventual purchase. The motivation being that the preferred type of offers (similar products vs. accessories) depend on whether a shopper is just window shopping or in a purchase process. External Posting Date: August 2, 2016 [Fulltext] Internal Posting Date: August 2, 2016 [Fulltext] Copyright 2016 Hewlett Packard Enterprise Development LP

2 User Behavior Recovery via Hidden Markov Models Analysis Alina Maor and Doron Shaked Software & Analytics Lab Hewlett Packard Labs July 20, 2016 Abstract In this report, we propose a method for user-behavior profiling and user-intention prediction based on Hidden Markov Models. User behavior is described by a sequence of user-actions. Note that there is no strict correspondence assumption between action order and user purpose or business process. On the contrary, we assume actions characterizing a purpose or a business process will have order variations or be interleaved with actions emerging from different purposes. The statistical nature of user behavior and the partial order assumption fit the statistical Hidden Markov Model. Different models are trained for different user profiles or to predict predetermined behavior types. Prediction can be incorporated with other predictors in a maximum likelihood setting. The proposed technique is generic and can be applied to profiling of any given number of (predefined) profiles. The only assumption is that profiles are characterized by different dynamics. The method was successfully applied in a two-group scenario to actions on HP.COM for predicting eventual purchase. The motivation being that the preferred type of offers (similar products vs. accessories) depend on whether a shopper is just window shopping or in a purchase process. Problem statement One of the most intriguing problems in multiuser applications analysis is timely profiling (classification) of application users. By user profiling we refer to the conceptual purpose or business process performed by users. In on-line shopping, examples of such profiles can be windowshopping, price comparison and purchase. In IT management 1

3 applications, examples of such profiles can be system initialization, code-update, commit and so on. The business implication of such profiling is enormous, providing: 1) application metering and monitoring via changes in users behavior; 2) application design efficiency; and 3) user-customized application experience. Henceforth, we concentrate on one specific example, namely profiling window shoppers and buyers in on-line shopping, focusing on HP.COM, but it should be clear that the proposed approach is applicable to other multi-user applications. User behavior is represented as a click-stream. The core problem our model solves is that user-actions are weakly unordered cluttered and interleaved: e.g., a purchase confirmation can be interrupted by additional browsing, pop-ups or sporadic actions such as an effort to contact the site administrator. As the first step to user-behavior analysis, we wish to classify users into two groups - potential buyers and window-shoppers, based on as little number of clicks as possible. This task is highly non-trivial due to the above mentioned complex nature of users-clicks. In addition the HP.COM site has additional complications including 1) HP.COM has a complex internal structure and URLs have functional duplicity such that identical user operations of different users often result in completely different click-streams. 2)A 30-minutes session termination rule. Our solution We base our solution on the long-known mechanism of Hidden Markov Models (HMMs), see [1] for details. HMM is a statistical model that assumes that there is a process generating observed outcomes (outputs). In our case, the outcomes are the user clicks (URLs). The process is governed by unknown unobserved, i.e., hidden, states. For on-line shopping, the states may refer to the implicit process states, e.g., looking for a particular product, comparing, or completing purchase. These states are defined by another hidden parameter - the user is either a potential buyer or a window-shopper. In our case, this hidden parameter is the profile type we want to uncover. Assume that a system has N hidden states and K possible outputs. By HMM, all hidden states are connected with a zero-memory Markov chain. When a system is in a certain state, there is a probability of generating every possible system output. Each output is dependent only on the current system-state and not on previous/future states or outputs. HMM definition includes also the initial probabilities of starting the system in each of the hidden states. For each user profile we train a separate HMM. For statistic validity, training and test data-sets are of the same size. The choice of N is a 2

4 difficult problem. Methods we developed based on randomization and divergence convergence indicated that N = 6 will be a good choice in this case. The convergence of HMM parameters is performed via the Baum-Welch method [1]. We regularize and normalize the model to account for rare outputs in the training sets outputs. Note that since K > 2500, rare outputs are quit common. Having trained two HMMs one for buying and one for non-buying, sequences are matched to both HMMs resulting in two likelihood values for buying or non-buying intentions. We perform two tests. The first scheme compares the likelihood values of and deciding whether the user is a buyer or not according to the higher likelihood. While this scheme is simple, it compares well against the Word-Based Algorithm (WBA). WBA is based on choosing a descriptive key-word(s) describing the desired user action and declaring a certain user behavior based on presence/absence of that key-word(s). For purchase analysis, two natural choices for such key-words are action-sets related to Cart, or Check-Out. The HMM approach always out-performs for short-medium sessions, and often for long sessions. We propose a novel algorithm, combining the advantages of both the WBA and the HMMbased approaches: Since both groups of users may perform the same sets of actions, the required correction to the HMM is the compensation for this noise. The accommodation for the noise will be both in offset and scaling - broadly speaking, buyer is declared when: LL buy αll not buy + β1 W BA + γ 0, (1) where LL is the logarithm of the likelihood, α is the LL scale, 1 W BA is the indicator function of word existence with weight β and γ is the regularization constant. Specifically, α emphasizes the difference in decisions performed by HMM and should not be too large, otherwise statistic validity of the decision will be violated. The value of γ provides constant offset compensating for noise in LL values due to high dimensionality. Finally, β is the tradeoff in decision made by HMM Vs WBA: The larger the value of β, the lesser the impact of HMM. Setting β, we degenerate to WBA. By taking β and γ to 0 and setting α = 1, we degenerate to the decision based solely on HMM. We omit further discussion on possible variation of the decision rule due to the abstract-length limitation. Evidence the solution works Data-Set: The HP.COM data-set contains a sample of session collected over a 6 month period and includes and sessions that were tagged as buyers/non-buyers, respectively, by an HP.COM coordinator, according to at least one purchase performed during the 3

5 last 2 months of the period. The number of sessions containing at least 5 clicks is buyers and 8559 non-buyers. For statistic validity, we create train-sets by randomization of 7000 buyer and non-buyer sessions out of the pool and use sessions for tests. The dimensionality of large sites is huge - in the HP.COM sample there are different URLs. Classified to K = types according to functionality (not removing functional duplicity). Limiting K to outputs appearing in the data sets more than twice reduced K to 4922 for the buyers model and 2671 for the non-buyers model. Some outputs are frequent and statistically valid while others are rare. We could omit these rare outputs, but we kept all. Starting with the first approach, we train two HMMs, and pause to address a difference in obtained behavior states: The HMM of buyers contains a clear purchasing state and other targeted browsing activities, while in case of non-buyers, purchasing activities are interleaved in almost all states with varying probability of occurrence. We proceed as following: 1) For every click in new sessions run the clicks through both HMMs and obtain both likelihood values. 2) Perform tests with these likelihood values: The simplest scheme compares the likelihoods and decides buyer/non-buyer by the higher value. The obtained results are generally better than WBA schemes as illustrated in Fig. 1. The x-axis is the number of clicks performed by user and the y-axis is accuracy, or the percentage of correct decisions. The four graphs represent four schemes: Magenta for the random-decision (RD), Red for the above HMM based decision, and Green and Blue for WBA scheme: users invoking Cart (Green) or Check-Out (Blue) are henceforth classified as buyers. As one can see, the accuracy of HMM is much higher than that of WBA and RD. Notice that the key advantage in users profiling comes from the earlier clicks. Next, we enrich the HMM method and the WBA schemes with the hybrid combination of both. The results of randomizing α, β and γ are demonstrated in Fig. 2 for decisions made at clicks 15 and 40: The x-axis and the y-axis are the recall and the precision percentage, respectively and the black graph is the break-even line. As we can see, for different values of α, β and γ the red-lines (matching different randomization values of α, β and γ) are closer to the upper rightmost corner of the plot, indicating the superiority of our hybrid combination. Competitive approaches There were numerous attempts to tackle the problem of online browsing modeling - see [2] and references therein. The main differences between the works that we are aware of and our setting is the sessions sub-division, unavailability of temporal data, and the decision making 4

6 algorithm (operating HMM per click) which is proposed herein. Current status and next steps We have developed a new tool for business intelligence analysis for the HP.COM analytic group. With slight changes, we also tested our approach on HP-SW Quality Center application focusing on profiling conceptual users-flow and flow-steps, rather than per-user categorization. We believe the HMM based tool is generically applicable to many applications and is suitable for different types of analysis. The attractive features of this solution is its implementation simplicity. The scheme can be plug-and-played in any system whose dynamics intrinsically assume hidden operational states. Specifically, HMMs-based analysis can be used for flow monitoring, system inspection and classification and we hope to integrate our scheme into an analytic platform. References [1] L.R. Rabiner, A Tutorial On Hidden Markov Models And Selected Applications In Speech Recognition, Proc. of the IEEE, vol. 77, no. 2, pp , Feb [2] A. L. Montgomery, S. Li, K. Srinivasan and J. C. Liechty, Modeling Online Browsing and Path Analysis Using Clickstream Data, Marketing Science, vol. 23, no. 4, pp , Fall

7 Figure 1: Accuracy of prediction of buying/non-buying intentions. Figure 2: Recall-Precision for non-buying intentions as function of Alpha, Beta and Gamma. 6

Evaluating Workflow Trust using Hidden Markov Modeling and Provenance Data

Evaluating Workflow Trust using Hidden Markov Modeling and Provenance Data Evaluating Workflow Trust using Hidden Markov Modeling and Provenance Data Mahsa Naseri and Simone A. Ludwig Abstract In service-oriented environments, services with different functionalities are combined

More information

Learning User Real-Time Intent for Optimal Dynamic Webpage Transformation

Learning User Real-Time Intent for Optimal Dynamic Webpage Transformation Learning User Real-Time Intent for Optimal Dynamic Webpage Transformation Amy Wenxuan Ding, Shibo Li and Patrali Chatterjee Web Appendix: Additional Results for the Two-State Proposed Model A. Program

More information

March 9, Hidden Markov Models and. BioInformatics, Part I. Steven R. Dunbar. Intro. BioInformatics Problem. Hidden Markov.

March 9, Hidden Markov Models and. BioInformatics, Part I. Steven R. Dunbar. Intro. BioInformatics Problem. Hidden Markov. and, and, March 9, 2017 1 / 30 Outline and, 1 2 3 4 2 / 30 Background and, Prof E. Moriyama (SBS) has a Seminar SBS, Math, Computer Science, Statistics Extensive use of program "HMMer" Britney (Hinds)

More information

Is Machine Learning the future of the Business Intelligence?

Is Machine Learning the future of the Business Intelligence? Is Machine Learning the future of the Business Intelligence Fernando IAFRATE : Sr Manager of the BI domain Fernando.iafrate@disney.com Tel : 33 (0)1 64 74 59 81 Mobile : 33 (0)6 81 97 14 26 What is Business

More information

Predicting Purchase Behavior of E-commerce Customer, One-stage or Two-stage?

Predicting Purchase Behavior of E-commerce Customer, One-stage or Two-stage? 2016 International Conference on Artificial Intelligence and Computer Science (AICS 2016) ISBN: 978-1-60595-411-0 Predicting Purchase Behavior of E-commerce Customer, One-stage or Two-stage? Chen CHEN

More information

Brochure Agility unlimited

Brochure Agility unlimited Brochure Agility unlimited Capgemini s Extreme Applications for Retail built on HPE ConvergedSystem for SAP HANA Brochure Page 2 Increase sales, improve margins, and optimize customer loyalty by leveraging

More information

Visual Mining Business Service Using Pixel Bar Charts

Visual Mining Business Service Using Pixel Bar Charts Visual Mining Business Service Using Pixel Bar Charts Ming C. Hao, Umeshwar Dayal, Fabio Casati Intelligent Enterprise Technologies Laboratory HP Laboratories Palo Alto HPL-2004-112 June 29, 2004* E-mail:

More information

ABA English: An elearning Leader Learns from Its Users E B C D A

ABA English: An elearning Leader Learns from Its Users E B C D A ABA English: An elearning Leader Learns from Its Users E E B C D A A ABA English is an online, subscription-based, distance learning platform that helps adults in over 7 countries learn English. ABA faced

More information

Digital Intelligence Solutions

Digital Intelligence Solutions WELCOME! You are about to enter the world of the Analytics Suite. First of all, thank you and welcome! In this document, we ll present an overview of the Analytics Suite applications available to you,

More information

Limits of Software Reuse

Limits of Software Reuse Technical Note Issued: 07/2006 Limits of Software Reuse L. Holenderski Philips Research Eindhoven c Koninklijke Philips Electronics N.V. 2006 Authors address: L. Holenderski WDC3-044; leszek.holenderski@philips.com

More information

SIMULATION METHOD FOR RELIABILITY ASSESSMENT OF ELECTRICAL DISTRIBUTION SYSTEMS

SIMULATION METHOD FOR RELIABILITY ASSESSMENT OF ELECTRICAL DISTRIBUTION SYSTEMS SIMULATION METHOD FOR RELIABILITY ASSESSMENT OF ELECTRICAL DISTRIBUTION SYSTEMS Johan Setréus, Lina Bertling, Shima Mousavi Gargari KTH, School of Electrical Engineering, Stockholm, Sweden johan.setreus@ee.kth.se

More information

Uncover possibilities with predictive analytics

Uncover possibilities with predictive analytics IBM Analytics Feature Guide IBM SPSS Modeler Uncover possibilities with predictive analytics Unlock the value of data you re already collecting by extracting information that opens a window into customer

More information

Speech Analytics Transcription Accuracy

Speech Analytics Transcription Accuracy Speech Analytics Transcription Accuracy Understanding Verint s speech analytics transcription and categorization accuracy Verint.com Twitter.com/verint Facebook.com/verint Blog.verint.com Table of Contents

More information

Tutorial Segmentation and Classification

Tutorial Segmentation and Classification MARKETING ENGINEERING FOR EXCEL TUTORIAL VERSION v171025 Tutorial Segmentation and Classification Marketing Engineering for Excel is a Microsoft Excel add-in. The software runs from within Microsoft Excel

More information

Gain control over all enterprise content

Gain control over all enterprise content Brochure Gain control over all enterprise content HP ControlPoint A better way to manage big data Most organizations today store data in a number of business systems and information repositories. This

More information

The application of hidden markov model in building genetic regulatory network

The application of hidden markov model in building genetic regulatory network J. Biomedical Science and Engineering, 2010, 3, 633-637 doi:10.4236/bise.2010.36086 Published Online June 2010 (http://www.scirp.org/ournal/bise/). The application of hidden markov model in building genetic

More information

RECOGNIZING USER INTENTIONS IN REAL-TIME

RECOGNIZING USER INTENTIONS IN REAL-TIME WHITE PAPER SERIES IPERCEPTIONS ACTIVE RECOGNITION TECHNOLOGY: RECOGNIZING USER INTENTIONS IN REAL-TIME Written by: Lane Cochrane, Vice President of Research at iperceptions Dr Matthew Butler PhD, Senior

More information

The Multi-touch marketing channel attribution problem - Can R help solve it? Dr Abigail Lebrecht

The Multi-touch marketing channel attribution problem - Can R help solve it? Dr Abigail Lebrecht The Multi-touch marketing channel attribution problem - Can R help solve it? Dr Abigail Lebrecht uswitch @A_Lebrecht PPC SEO Direct Buy Banner SEO Buy SEO Direct The Attribution Problem Customers are coming

More information

Analytics with Intelligent Edge Sergey Serebryakov Research Engineer IEEE COMMUNICATIONS SOCIETY

Analytics with Intelligent Edge Sergey Serebryakov Research Engineer IEEE COMMUNICATIONS SOCIETY Analytics with Intelligent Edge Sergey Serebryakov Research Engineer IEEE COMMUNICATIONS SOCIETY Deep learning today Vision Speech Text Other Search & information extraction Security/Video surveillance

More information

Getting Started with OptQuest

Getting Started with OptQuest Getting Started with OptQuest What OptQuest does Futura Apartments model example Portfolio Allocation model example Defining decision variables in Crystal Ball Running OptQuest Specifying decision variable

More information

Research on Problems and Countermeasures of Social E-commerce Operation

Research on Problems and Countermeasures of Social E-commerce Operation 2018 7th International Conference on Social Science, Education and Humanities Research (SSEHR 2018) Research on Problems and Countermeasures of Social E-commerce Operation Qiang Jin Xijing University,

More information

Addressing Predictive Maintenance with SAP Predictive Analytics

Addressing Predictive Maintenance with SAP Predictive Analytics SAP Predictive Analytics Addressing Predictive Maintenance with SAP Predictive Analytics Table of Contents 2 Extract Predictive Insights from the Internet of Things 2 Design and Visualize Predictive Models

More information

Hidden Markov Models. Some applications in bioinformatics

Hidden Markov Models. Some applications in bioinformatics Hidden Markov Models Some applications in bioinformatics Hidden Markov models Developed in speech recognition in the late 1960s... A HMM M (with start- and end-states) defines a regular language L M of

More information

Using SAP with HP Virtualization and Partitioning

Using SAP with HP Virtualization and Partitioning Using SAP with HP Virtualization and Partitioning Introduction... 2 Overview of Virtualization and Partitioning Technologies... 2 Physical Servers... 2 Hard Partitions npars... 3 Virtual Partitions vpars...

More information

Creating a CMD in CDX

Creating a CMD in CDX Creating a CMD in CDX Part 5 - Supplier Submissions June, 2016 The CDX User Manual contains concise information explaining the CDX system, yet assumes familiarity with the industry principles involved.

More information

Adobe and Hadoop Integration

Adobe and Hadoop Integration Predictive Behavioral Analytics Adobe and Hadoop Integration JANUARY 2016 SYNTASA Copyright 1.0 Introduction For many years large enterprises have relied on the Adobe Marketing Cloud for capturing and

More information

A stochastic approach to multi channel attribution in online advertising

A stochastic approach to multi channel attribution in online advertising A stochastic approach to multi channel attribution in online advertising Authors Venkatesh Sreedharan (venkateshw@freshworks.com Swaminathan Padmanabhan (swami@freshworks.com Machine Learning @Freshworks

More information

ProGen: GPHMM for prokaryotic genomes

ProGen: GPHMM for prokaryotic genomes ProGen: GPHMM for prokaryotic genomes Sharad Akshar Punuganti May 10, 2011 Abstract ProGen is an implementation of a Generalized Pair Hidden Markov Model (GPHMM), a model which can be used to perform both

More information

Determining presence/absence threshold for your dataset

Determining presence/absence threshold for your dataset Determining presence/absence threshold for your dataset In PanCGHweb there are two ways to determine the presence/absence calling threshold. One is based on Receiver Operating Curves (ROC) generated for

More information

HP Operations Analytics

HP Operations Analytics HP Operations Analytics March 10, 2015 Joe Hasenohr, Ryan Huneidi HP Software HP Business Service Management Ensure optimal IT performance and availability in a dynamic world Business Service Management

More information

Optimization of Click-Through Rate Prediction in the Yandex Search Engine

Optimization of Click-Through Rate Prediction in the Yandex Search Engine ISSN 5-155, Automatic Documentation and Mathematical Linguistics, 213, Vol. 47, No. 2, pp. 52 58. Allerton Press, Inc., 213. Original Russian Text K.E. Bauman, A.N. Kornetova, V.A. Topinskii, D.A. Khakimova,

More information

INSIGHTS. Driving Decisions With Data: iquanti s Hybrid Approach to Attribution Modeling. Ajay Rama, Pushpendra Kumar

INSIGHTS. Driving Decisions With Data: iquanti s Hybrid Approach to Attribution Modeling. Ajay Rama, Pushpendra Kumar INSIGHTS Driving Decisions With Data: iquanti s Hybrid Approach to Attribution Modeling Ajay Rama, Pushpendra Kumar TABLE OF CONTENTS Introduction The Marketer s Dilemma 1. Media Mix Modeling (MMM) 1.

More information

ACC: Review of the Weekly Compensation Duration Model

ACC: Review of the Weekly Compensation Duration Model ACC: Review of the Weekly Compensation Duration Model Prepared By: Knoware Creation Date July 2013 Version: 1.0 Author: M. Kelly Mara Knoware +64 21 979 799 kelly.mara@knoware.co.nz Document Control Change

More information

THE ECOMMERCE MARKETER'S GUIDE TO ADVANCED AUDIENCE TARGETING

THE ECOMMERCE MARKETER'S GUIDE TO ADVANCED AUDIENCE TARGETING THE ECOMMERCE MARKETER'S GUIDE TO ADVANCED AUDIENCE TARGETING Why, when, and how to improve your store conversion rates with targeted on-site displays that reach the right person with the right message

More information

A Protein Secondary Structure Prediction Method Based on BP Neural Network Ru-xi YIN, Li-zhen LIU*, Wei SONG, Xin-lei ZHAO and Chao DU

A Protein Secondary Structure Prediction Method Based on BP Neural Network Ru-xi YIN, Li-zhen LIU*, Wei SONG, Xin-lei ZHAO and Chao DU 2017 2nd International Conference on Artificial Intelligence: Techniques and Applications (AITA 2017 ISBN: 978-1-60595-491-2 A Protein Secondary Structure Prediction Method Based on BP Neural Network Ru-xi

More information

Chapter Analytical Tool and Setting Parameters: determining the existence of differences among several population means

Chapter Analytical Tool and Setting Parameters: determining the existence of differences among several population means Chapter-6 Websites Services Quality Analysis of Data & Results 6.1 Analytical Tool and Setting Parameters: 1. Analysis of Variance (ANOVA): I have shown a relative scenario between his results in the results

More information

3 Ways to Improve Your Targeted Marketing with Analytics

3 Ways to Improve Your Targeted Marketing with Analytics 3 Ways to Improve Your Targeted Marketing with Analytics Introduction Targeted marketing is a simple concept, but a key element in a marketing strategy. The goal is to identify the potential customers

More information

Genetic Algorithm for Predicting Protein Folding in the 2D HP Model

Genetic Algorithm for Predicting Protein Folding in the 2D HP Model Genetic Algorithm for Predicting Protein Folding in the 2D HP Model A Parameter Tuning Case Study Eyal Halm Leiden Institute of Advanced Computer Science, University of Leiden Niels Bohrweg 1 2333 CA Leiden,

More information

Traffic Safety Measures Using Multiple Streams Real Time Data

Traffic Safety Measures Using Multiple Streams Real Time Data CAIT-UTC-055 Traffic Safety Measures Using Multiple Streams Real Time Data FINAL REPORT January 2017 Submitted by: Mohsen Jafari Professor of Industrial and Systems Engineering Center for Advanced Infrastructure

More information

Constrained Hidden Markov Models for Population-based Haplotyping

Constrained Hidden Markov Models for Population-based Haplotyping Constrained Hidden Markov Models for Population-based Haplotyping Extended abstract Niels Landwehr, Taneli Mielikäinen 2, Lauri Eronen 2, Hannu Toivonen,2, and Heikki Mannila 2 Machine Learning Lab, Dept.

More information

TEXT MINING APPROACH TO EXTRACT KNOWLEDGE FROM SOCIAL MEDIA DATA TO ENHANCE BUSINESS INTELLIGENCE

TEXT MINING APPROACH TO EXTRACT KNOWLEDGE FROM SOCIAL MEDIA DATA TO ENHANCE BUSINESS INTELLIGENCE International Journal of Advance Research In Science And Engineering http://www.ijarse.com TEXT MINING APPROACH TO EXTRACT KNOWLEDGE FROM SOCIAL MEDIA DATA TO ENHANCE BUSINESS INTELLIGENCE R. Jayanthi

More information

Bringing Criteo s Performance to Search Engine Marketing

Bringing Criteo s Performance to Search Engine Marketing Jason Lehmbeck, GM Search Investor Day, September 2016 Bringing Criteo s Performance to Search Engine Marketing Safe Harbor Statement This presentation contains forward-looking statements that are based

More information

Multidimensional Aptitude Battery-II (MAB-II) Extended Report

Multidimensional Aptitude Battery-II (MAB-II) Extended Report Multidimensional Aptitude Battery-II (MAB-II) Extended Report Name: Sam Sample A g e : 30 (Age Group 25-34) Gender: Male Report Date: January 17, 2017 The profile and report below are based upon your responses

More information

The Future of Workload Automation in the Application Economy

The Future of Workload Automation in the Application Economy The Future of Workload Automation in the Application Economy Success Requires Agility in the Application Economy The link between data center operations and business agility has never been stronger. If

More information

Finding Hidden Intelligence with Predictive Analysis of Data Mining

Finding Hidden Intelligence with Predictive Analysis of Data Mining Finding Hidden Intelligence with Predictive Analysis of Data Mining Rafal Lukawiecki Strategic Consultant, Project Botticelli Ltd rafal@projectbotticelli.com Objectives Show use of Microsoft SQL Server

More information

Analysis and Modelling of Flexible Manufacturing System

Analysis and Modelling of Flexible Manufacturing System Analysis and Modelling of Flexible Manufacturing System Swetapadma Mishra 1, Biswabihari Rath 2, Aravind Tripathy 3 1,2,3Gandhi Institute For Technology,Bhubaneswar, Odisha, India --------------------------------------------------------------------***----------------------------------------------------------------------

More information

Adobe and Hadoop Integration

Adobe and Hadoop Integration Predictive Behavioral Analytics Adobe and Hadoop Integration DECEMBER 2016 SYNTASA Copyright 1.0 Introduction For many years large enterprises have relied on the Adobe Marketing Cloud for capturing and

More information

Correlating instrumentation data to system states: A building block for automated diagnosis and control

Correlating instrumentation data to system states: A building block for automated diagnosis and control Correlating instrumentation data to system states: A building block for automated diagnosis and control Ira Cohen Ira.Cohen@hp.com Joint work with: Moises Goldszmidt, Terence Kelly, and Julie Symons of

More information

Analytic Alphabet Soup: IoT, AI & ESP Big Data Analytics is a game changer in our Connected World

Analytic Alphabet Soup: IoT, AI & ESP Big Data Analytics is a game changer in our Connected World Analytic Alphabet Soup: IoT, AI & ESP Big Data Analytics is a game changer in our Connected World Eric Hunley Director Enterprise Solutions SAS Great leaders are almost always great simplifiers, who can

More information

Computational Sustainability: Smart Buildings

Computational Sustainability: Smart Buildings Computational Sustainability: Smart Buildings CS 325: Topics in Computational Sustainability, Spring 2016 Manish Marwah Senior Research Scientist Hewlett Packard Labs manish.marwah@hpe.com Building Energy

More information

Provided By WealthyAffiliate.com

Provided By WealthyAffiliate.com The Content within this guide is protected by Copyright and is the sole property of Kyle & Carson of Niche Marketing Inc. This guide may not be modified, copied, or sold. You may however, give this guide

More information

2 Maria Carolina Monard and Gustavo E. A. P. A. Batista

2 Maria Carolina Monard and Gustavo E. A. P. A. Batista Graphical Methods for Classifier Performance Evaluation Maria Carolina Monard and Gustavo E. A. P. A. Batista University of São Paulo USP Institute of Mathematics and Computer Science ICMC Department of

More information

Modelling buyer behaviour/3 How survey results can mislead

Modelling buyer behaviour/3 How survey results can mislead Publishing Date: June 1993. 1993. All rights reserved. Copyright rests with the author. No part of this article may be reproduced without written permission from the author. Modelling buyer behaviour/3

More information

Manufacturing Technology Committee Risk Management Working Group Risk Management Training Guides

Manufacturing Technology Committee Risk Management Working Group Risk Management Training Guides Manufacturing Technology Committee Management Working Group Management Training Guides Ranking and Filtering 1 Overview Ranking and Filtering is one of the most common facilitation methods used for Management.

More information

Analysing Clickstream Data: From Anomaly Detection to Visitor Profiling

Analysing Clickstream Data: From Anomaly Detection to Visitor Profiling Analysing Clickstream Data: From Anomaly Detection to Visitor Profiling Peter I. Hofgesang and Wojtek Kowalczyk Free University of Amsterdam, Department of Computer Science, Amsterdam, The Netherlands

More information

Inferring Social Ties across Heterogeneous Networks

Inferring Social Ties across Heterogeneous Networks Inferring Social Ties across Heterogeneous Networks CS 6001 Complex Network Structures HARISH ANANDAN Introduction Social Ties Information carrying connections between people It can be: Strong, weak or

More information

THREE LEVEL HIERARCHICAL BAYESIAN ESTIMATION IN CONJOINT PROCESS

THREE LEVEL HIERARCHICAL BAYESIAN ESTIMATION IN CONJOINT PROCESS Please cite this article as: Paweł Kopciuszewski, Three level hierarchical Bayesian estimation in conjoint process, Scientific Research of the Institute of Mathematics and Computer Science, 2006, Volume

More information

Application Note. Operating Micropumps at Low Flow Rates

Application Note. Operating Micropumps at Low Flow Rates Application Note Operating Micropumps at Low Flow Rates In the following application note, achieving low flow rates with the micropumps of Bartels Mikrotechnik will be discussed. All formulas and values

More information

HP AppSystems for SAP HANA

HP AppSystems for SAP HANA HP AppSystems for Kostiantyn Grygortsov Technical consultant, HP Ukraine Copyright 2012 Hewlett-Packard Development Company, L.P. The Copyright information 2012 contained Hewlett-Packard herein is Development

More information

Segmentation Modeling

Segmentation Modeling Deepen Customer Understanding There is an infinitely wide gamut of human behavior, interests and characteristics to account for as a modern business operating in the online marketplace. To understand consumers,

More information

LEARN HOW LENOVO & Tesco MAKE ADOBE & ENTERPRISE DATA ACTIONABLE IN HADOOP

LEARN HOW LENOVO & Tesco MAKE ADOBE & ENTERPRISE DATA ACTIONABLE IN HADOOP LEARN HOW LENOVO & Tesco MAKE ADOBE & ENTERPRISE DATA ACTIONABLE IN HADOOP David Searle General Manager, EMEA @SyntasaCo Ashish Braganza Global Business Intelligence Director @Lenovo Simon Ricketts Data,

More information

R&D. R&D digest. by simulating different scenarios and counting their cost in profits and customer service.

R&D. R&D digest. by simulating different scenarios and counting their cost in profits and customer service. digest Managing a utility s assets: new software shows how Asset management is no longer just about using our people, factories, equipment and processes wisely. Just as importantly, it s about planning

More information

The Appliance Based Approach for IT Infrastructure Management

The Appliance Based Approach for IT Infrastructure Management The Appliance Based Approach for IT Infrastructure Management This white paper examines the key issues faced by IT managers in managing the IT infrastructure of their organizations. A new solution using

More information

Consumer Referral in a Small World Network

Consumer Referral in a Small World Network Consumer Referral in a Small World Network Tackseung Jun 1 Beom Jun Kim 2 Jeong-Yoo Kim 3 August 8, 2004 1 Department of Economics, Kyung Hee University, 1 Hoegidong, Dongdaemunku, Seoul, 130-701, Korea.

More information

Put cloud-based insights to work for your business

Put cloud-based insights to work for your business Put cloud-based insights to work for your business As the volume of data grows, businesses are using the power of the cloud to gather, analyze, and visualize data from internal and external sources to

More information

Drive IT and Business Improvements with In-Memory Computing

Drive IT and Business Improvements with In-Memory Computing Drive IT and Business Improvements with In-Memory Computing Pervinder Johar, Hewlett Packard Global Supply Chain Systems May 2013 Copyright 2012 2013 Hewlett-Packard Development Company, L.P. The information

More information

360 o View of the Customer. Managing Big Data. Partnering with IT. Strategic Analytics. Optimization Engines. Closing the Loop CRITICAL ENABLERS

360 o View of the Customer. Managing Big Data. Partnering with IT. Strategic Analytics. Optimization Engines. Closing the Loop CRITICAL ENABLERS 360 o View of the Customer Managing Big Data Partnering with IT Strategic Analytics Optimization Engines Closing the Loop CRITICAL ENABLERS 360 View Of Customer Continuous Learning Why Who What Where Customer

More information

Bionano Access 1.0 Software User Guide

Bionano Access 1.0 Software User Guide Bionano Access 1.0 Software User Guide Document Number: 30142 Document Revision: A For Research Use Only. Not for use in diagnostic procedures. Copyright 2017 Bionano Genomics, Inc. All Rights Reserved.

More information

Infor LN Procurement User's Guide for Purchase Requisition Workbench

Infor LN Procurement User's Guide for Purchase Requisition Workbench Infor LN Procurement User's Guide for Purchase Requisition Workbench Copyright 2018 Infor Important Notices The material contained in this publication (including any supplementary information) constitutes

More information

Learning-based Surgical Workflow Detection from Intra- Operative Signals

Learning-based Surgical Workflow Detection from Intra- Operative Signals Learning-based Surgical Workflow Detection from Intra- Operative Signals Ralf Stauder 1,*, Ergün Kayis 1,*, Nassir Navab 1,2 1 Computer Aided Medical Procedures, Technical University of Munich, Germany

More information

Getting Started with HLM 5. For Windows

Getting Started with HLM 5. For Windows For Windows Updated: August 2012 Table of Contents Section 1: Overview... 3 1.1 About this Document... 3 1.2 Introduction to HLM... 3 1.3 Accessing HLM... 3 1.4 Getting Help with HLM... 3 Section 2: Accessing

More information

NEXT GENERATION PREDICATIVE ANALYTICS USING HP DISTRIBUTED R

NEXT GENERATION PREDICATIVE ANALYTICS USING HP DISTRIBUTED R 1 A SOLUTION IS NEEDED THAT NOT ONLY HANDLES THE VOLUME OF BIG DATA OR HUGE DATA EASILY, BUT ALSO PRODUCES INSIGHTS INTO THIS DATA QUICKLY NEXT GENERATION PREDICATIVE ANALYTICS USING HP DISTRIBUTED R A

More information

SBW For Windows TM. Version Advanced Quoter Guide Module 7.2: Pricing

SBW For Windows TM. Version Advanced Quoter Guide Module 7.2: Pricing SBW For Windows TM Version 12.0 Advanced Quoter Guide Module 7.2: Pricing Updated June 2005 SBW for Windows 12.0 Page 2 of 7 June 2005 Copyright 2004 Hewlett-Packard Development Company, L.P. The information

More information

A New Analysis Concept in Applying Software Reliability Growth Models and Tool Implementation: The SafeMan

A New Analysis Concept in Applying Software Reliability Growth Models and Tool Implementation: The SafeMan Journal of Software Engineering and Applications, 2014, 7, 396-405 Published Online May 2014 in SciRes. http://www.scirp.org/journal/jsea http://dx.doi.org/10.4236/jsea.2014.75036 A New Analysis Concept

More information

Churn Prediction for Game Industry Based on Cohort Classification Ensemble

Churn Prediction for Game Industry Based on Cohort Classification Ensemble Churn Prediction for Game Industry Based on Cohort Classification Ensemble Evgenii Tsymbalov 1,2 1 National Research University Higher School of Economics, Moscow, Russia 2 Webgames, Moscow, Russia etsymbalov@gmail.com

More information

FAILURE ANALYSIS OF POWER TRANSFORMERS BY DGA, OIL TESTS AND MARKOV APPROACH

FAILURE ANALYSIS OF POWER TRANSFORMERS BY DGA, OIL TESTS AND MARKOV APPROACH FAILURE ALYSIS OF POWER TRANSFORMERS BY DGA, OIL TESTS AND MARKOV APPROACH Mohammed Abdul Rahman Uzair Research Scholar, Department of EEE, GITAM University, Hyderabad, Telangana, INDIA Dr. Basavaraja

More information

1. Can you explain the PDCA cycle and where testing fits in?

1. Can you explain the PDCA cycle and where testing fits in? 1. Can you explain the PDCA cycle and where testing fits in? Software testing is an important part of the software development process. In normal software development there are four important steps, also

More information

INTRODUCTION. It is the process used to identify the correctness, completeness and quality of developed computer software.

INTRODUCTION. It is the process used to identify the correctness, completeness and quality of developed computer software. INTRODUCTION It is the process used to identify the correctness, completeness and quality of developed computer software. It is the process of executing a program/application under positive and negative

More information

IMPACT OF DIGITAL MARKETING ON BUYER BEHAVIOR

IMPACT OF DIGITAL MARKETING ON BUYER BEHAVIOR IMPACT OF DIGITAL MARKETING ON BUYER BEHAVIOR RAJESH PERSHAD Assoc. Professor, MBA Dept, JBIET, Moinabad, Hyderabad. ABSTRACT Marketers have been able to provide superior value in creative ways that go

More information

CSE 255 Lecture 3. Data Mining and Predictive Analytics. Supervised learning Classification

CSE 255 Lecture 3. Data Mining and Predictive Analytics. Supervised learning Classification CSE 255 Lecture 3 Data Mining and Predictive Analytics Supervised learning Classification Last week Last week we started looking at supervised learning problems Last week We studied linear regression,

More information

How to drive profitability in uncertain times. How advanced analytics can help your business to become a quality-connected enterprise

How to drive profitability in uncertain times. How advanced analytics can help your business to become a quality-connected enterprise How to drive profitability in uncertain times How advanced analytics can help your business to become a quality-connected enterprise The importance of an end-to-end approach No one can predict the economic

More information

Step-by-Step Guide Creating Shopping Carts from Invoice Attached Forms

Step-by-Step Guide Creating Shopping Carts from Invoice Attached Forms In This Guide Selecting an Invoice Attached form Creating a shopping cart This guide demonstrates how to create an Invoice Attached form for goods or services. This form is used for payments for which

More information

Novel Method for Discrimination of Conserved Genes through Numerical Characterization of DNA Sequences

Novel Method for Discrimination of Conserved Genes through Numerical Characterization of DNA Sequences Internet Electronic Journal of Molecular Design 2003, 2, 000 000 Novel Method for Discrimination of Conserved Genes through Numerical Characterization of DNA Sequences Ashesh Nandy,* 1 Environmental Programme

More information

by Xindong Wu, Kui Yu, Hao Wang, Wei Ding

by Xindong Wu, Kui Yu, Hao Wang, Wei Ding Online Streaming Feature Selection by Xindong Wu, Kui Yu, Hao Wang, Wei Ding 1 Outline 1. Background and Motivation 2. Related Work 3. Notations and Definitions 4. Our Framework for Streaming Feature Selection

More information

HP Cloud Maps for rapid provisioning of infrastructure and applications

HP Cloud Maps for rapid provisioning of infrastructure and applications Technical white paper HP Cloud Maps for rapid provisioning of infrastructure and applications Table of contents Executive summary 2 Introduction 2 What is an HP Cloud Map? 3 HP Cloud Map components 3 Enabling

More information

Netflix Optimization: A Confluence of Metrics, Algorithms, and Experimentation. CIKM 2013, UEO Workshop Caitlin Smallwood

Netflix Optimization: A Confluence of Metrics, Algorithms, and Experimentation. CIKM 2013, UEO Workshop Caitlin Smallwood Netflix Optimization: A Confluence of Metrics, Algorithms, and Experimentation CIKM 2013, UEO Workshop Caitlin Smallwood 1 Allegheny Monongahela Ohio River 2 TV & Movie Enjoyment Made Easy Stream any video

More information

The impact of banner advertisement frequency on brand awareness

The impact of banner advertisement frequency on brand awareness The impact of banner advertisement frequency on brand awareness Author Hussain, Rahim, Sweeney, Arthur, Sullivan Mort, Gillian Published 2007 Conference Title 2007 ANZMAC Conference Proceedings Copyright

More information

Enterprise Document Assessment Methodology. Global 20 Oil & Gas Office Documents Costs

Enterprise Document Assessment Methodology. Global 20 Oil & Gas Office Documents Costs EDAM Enterprise Document Assessment Methodology Global 20 Oil & Gas The data represented in this report is summarized for simplicity. It is based on our independent assessment, research and computational

More information

IMPLEMENTATION FOR ENHANCING SECURITY OF RFID CARD

IMPLEMENTATION FOR ENHANCING SECURITY OF RFID CARD IMPLEMENTATION FOR ENHANCING SECURITY OF RFID CARD Shilpa S. Badhiye1,Prof.Rupali S. Khule2 1 student, Electronics and telecommunication Department, MCOERC, Maharashtra, India 2 Professor, Electronics

More information

CS273: Algorithms for Structure Handout # 5 and Motion in Biology Stanford University Tuesday, 13 April 2004

CS273: Algorithms for Structure Handout # 5 and Motion in Biology Stanford University Tuesday, 13 April 2004 CS273: Algorithms for Structure Handout # 5 and Motion in Biology Stanford University Tuesday, 13 April 2004 Lecture #5: 13 April 2004 Topics: Sequence motif identification Scribe: Samantha Chui 1 Introduction

More information

Certificate Program in Digital Marketing

Certificate Program in Digital Marketing Certificate Program in Digital Marketing Months 3 months Hours 72 hours Fees Rs. 40,000 all inclusive 1. Introduction to Digital Media & Fundamentals The media shift from traditional media to Digital Understanding

More information

PCA and SOM based Dimension Reduction Techniques for Quaternary Protein Structure Prediction

PCA and SOM based Dimension Reduction Techniques for Quaternary Protein Structure Prediction PCA and SOM based Dimension Reduction Techniques for Quaternary Protein Structure Prediction Sanyukta Chetia Department of Electronics and Communication Engineering, Gauhati University-781014, Guwahati,

More information

A Bayesian Predictor of Airline Class Seats Based on Multinomial Event Model

A Bayesian Predictor of Airline Class Seats Based on Multinomial Event Model 2016 IEEE International Conference on Big Data (Big Data) A Bayesian Predictor of Airline Class Seats Based on Multinomial Event Model Bingchuan Liu Ctrip.com Shanghai, China bcliu@ctrip.com Yudong Tan

More information

DECISION MAKING. Chapter - 4. B. H. Gardi College of Engineering & Technology, RAJKOT Department of Master of Computer Application

DECISION MAKING. Chapter - 4. B. H. Gardi College of Engineering & Technology, RAJKOT Department of Master of Computer Application Prepared By :- Mr. Ajay A. Ardeshana MCA Lecturer At GARDI VIDYAPITH RAJKOT. Email :- ajay.24021985@gmail.com Mobile :- + 91 95588 20298 Chapter - 4 DECISION MAKING Email : ajay.24021985@gmail.com Mobile

More information

CHAPTER 6 A CDMA BASED ANTI-COLLISION DETERMINISTIC ALGORITHM FOR RFID TAGS

CHAPTER 6 A CDMA BASED ANTI-COLLISION DETERMINISTIC ALGORITHM FOR RFID TAGS CHAPTER 6 A CDMA BASED ANTI-COLLISION DETERMINISTIC ALGORITHM FOR RFID TAGS 6.1 INTRODUCTION Applications making use of Radio Frequency Identification (RFID) technology with large tag populations often

More information

Comparison of Deterministic and Stochastic Production Planning. Approaches in Sawmills by Integrating Design of Experiments and Monte-

Comparison of Deterministic and Stochastic Production Planning. Approaches in Sawmills by Integrating Design of Experiments and Monte- Comparison of Deterministic and Stochastic Production Planning Approaches in Sawmills by Integrating Design of Experiments and Monte- Carlo simulation Naghmeh Vahidian A Thesis in the Department of Mechanical

More information

IBM Predictive Maintenance and Quality (Version 2.0) IBM Redbooks Solution Guide

IBM Predictive Maintenance and Quality (Version 2.0) IBM Redbooks Solution Guide IBM Predictive Maintenance and Quality (Version 2.0) IBM Redbooks Solution Guide The IBM Predictive Maintenance and Quality solution (PMQ) uses information collected about products, processes, and assets

More information

Finding Regularity in Protein Secondary Structures using a Cluster-based Genetic Algorithm

Finding Regularity in Protein Secondary Structures using a Cluster-based Genetic Algorithm Finding Regularity in Protein Secondary Structures using a Cluster-based Genetic Algorithm Yen-Wei Chu 1,3, Chuen-Tsai Sun 3, Chung-Yuan Huang 2,3 1) Department of Information Management 2) Department

More information

Similarity-Based Sampling: Testing a Model of Price Psychophysics

Similarity-Based Sampling: Testing a Model of Price Psychophysics Similarity-Based Sampling: Testing a Model of Price Psychophysics Jing Qian (j.qian@warwick.ac.uk) Department of Psychology, University of Warwick Coventry, CV4 7AL, UK Gordon D.A. Brown (g.d.a.brown@warwick.ac.uk)

More information

AiM User Guide Inventory Management Module

AiM User Guide Inventory Management Module Inventory Management Module 2009 AssetWorks Inc. 1777 NE Loop 410, Suite 1250 San Antonio, Texas 78217 (800) 268-0325 Table of Contents AiM User Guide INTRODUCTION... 7 CHAPTERS... 7 PART 1... 7 PART 2...

More information