Recommendation techniques and end-user guidance for collaborative Knowledge-Intensive Processes Sebastian Fröhlich 18.05.2015 Software Engineering for Business Information Systems (sebis) Department of Informatics Technische Universität München, Germany wwwmatthes.in.tum.de
Agenda 1. Introduction 1. Motivation 2. Research Questions 2. Use Cases 3. Approach 1. Conceptual Structure 2. Related Work 3. Conceptual Approach 4. Roadmap 5. Discussion Initial Presentation Master Thesis Sebastian Fröhlich 2
Adapted from current Darwin feed Introduction - Motivation Case A Case B Seitenfenster ist offen Benutzer öffnet Seitenfenster Mitarbeiter 1 Create mockups Identify relevant literature Make mockups Perform literature study Relevant Nuter gibt Namen des ein Nutzer ermittelt mögliche Attribute Attribute ermittelt Attribute werden angelegt Task angelegt V Nutzer ermittelt mögliche Expertisen Expertisen ermittelt Nutzer prüft Unterlagen nach möglicher Dauer Mögliche Dauer gefunden Nutzer gibt Sart/Enddatum ein Create second iterate of mockups Improve mockups Attribute sind angelegt Attribute werden Task zugewiesen Expertisen werden hinzugefügt Task enthält geeignete Expertisen Start/Enddatum eingegeben Attribute sind zugewiesen V... Guidance and recommendation Structuring (similar naming lead to higher recognition) Usability (recommend only relevant options) Efficiency (reuse of predefined Structure with one click) Initial Presentation Master Thesis Sebastian Fröhlich 3
Introduction - Research Questions How can recommendations help to guide users? Which existing approaches can be applied to achieve the recommendation? How do a recommendation model look like? Initial Presentation Master Thesis Sebastian Fröhlich 4
Use Cases (1): Task Recommendations Current State Simple Task Recommendations Advanced Task Recommendations Initial Presentation Master Thesis Sebastian Fröhlich 5
Use Cases (2): Expertise Recommendations Current State Simple Expertise Recommendations Advanced Expertise Recommendations Initial Presentation Master Thesis Sebastian Fröhlich 6
Use Cases (3): Attribute Recommendations Current State Simple Attribute Recommendations Advanced Attribute Recommendations Initial Presentation Master Thesis Sebastian Fröhlich 7
Subpage Printable Subpage Checklist Approach - Conceptual Structure Page Hierarchy Conceptual Structure Conceptual Element Company Party Organisation Preparation Company Party Invitation Management Agenda Topic Entertainment Games Preparation Invitation Management Agenda Wrap Up Agenda Create Topic Make entertainment Programm Speech Wrap Up Calculation Topic Entertainment Initial Presentation Master Thesis Sebastian Fröhlich 8
Approach - Related Work (1) Context insensitive Recommendations (Nauerz, 2012) + Important at any time - Can be completely unnecessary in context - Maybe false point of time Task A: #12 Task B: #4 Task B: #3 Task A Clustering (Jannach, 2011) (Klahold, 2009) + High chance that similar users do similar tasks - Can be completely unnecessary in context Nauerz, Andreas (2012): Adapting and recommending content and expertise in highly collaborative web portals. Univ., Diss.--Jena, 2012. München: Dr. Hut (Informatik). Jannach, Dietmar (2011): Recommender systems. An introduction. New York: Cambridge University Press. Klahold, André (2009): Empfehlungssysteme. Recommender Systems ; Grundlagen, Konzepte und Lösungen. 1. Aufl. Wiesbaden: Vieweg + Teubner (Aus dem Programm IT-Management und -Anwendungen). Initial Presentation Master Thesis Sebastian Fröhlich 9
Approach - Related Work (2) Based on History: By similarity (Motahari-Nezhad et al., 2011) + Similar finished tasks in similar order => high chance to predict next task in context - Only good result if similarity is high enough Own: A Existing B A B C B A D C User / Crowd based (Dorn et al., 2010) + Adepts better to completed workflows - No direct support for different task stages With heuristics (van der Aalst et al., 2003) + Can deal with noise and incomplete logs - Different workflows lead to very large table with mostly null values Motahari-Nezhad, Hamid Reza; Bartolini, Claudio; Graupner, Sven; Spence, Susan (2012): Adaptive Case Management in the Social Enterprise. In : Service-oriented computing. 10th international conference, ICSOC 2012, Shanghai, China, November 12-15, 2012, proceedings. Heidelberg, New York: Springer (LNCS sublibrary. SL 2, Programming and software engineering, 7636), pp. 550 557. Dorn, Christoph; Burkhart, Thomas; Werth, Dirk; Dustdar, Schahram: Self-adjusting Recommendations for People-driven Ad-hoc Processes. van der Aalst, W.M.P.; van Dongen, B. F.; Herbst, J.; Maruster, L.; Schimm, G.; Weijters, A.J.M.M. (2003): Workflow mining: A survey of issues and approaches. In Data & Knowledge Engineering 47 (2), pp. 237 267. DOI: 10.1016/S0169-023X(03)00066-1. Initial Presentation Master Thesis Sebastian Fröhlich 10
Approach - Conceptional Model (1): Task Co mpan y Party Instance 1 Instance 2 Prep aration Invitation Man agemen t # Agenda Wrap Up Cat ering Topic Enter tainmen t Preparation Organize Ensure Brief colleagues Determine Audience Determine Differences Organize Ensure Preparation Brief colleagues Order Choose food Find possible s Ensure Order Choose food Find possible s Ensure filter and sort Initial Presentation Master Thesis Sebastian Fröhlich 11
Executor Overall Cost Executor Approach - Conceptional Model (2): Attribute Instance 1 Instance 2 Preparation Determine Differences Preparation filter and sort Initial Presentation Master Thesis Sebastian Fröhlich 12
Approach - Conceptional Model (3): Expertise Preparation Organize Ensure Scheduling Order Choose food Scheduling Gourmet Use Expertises of higher level Task(s) s Find possible s Ensure Use Expertise of other instances Initial Presentation Master Thesis Sebastian Fröhlich 13
Roadmap Mar Apr May Jun Jul Aug Sep Oct Nov Literature Review Conceptional Design Implementation Evaluation Writing the thesis Begin End Initial Presentation Master Thesis Sebastian Fröhlich 14
Discussion Sebastian Fröhlich Technische Universität München Department of Informatics Chair of Software Engineering for Business Information Systems Boltzmannstraße 3 85748 Garching bei München Tel +49.89.289. 17129 Fax +49.89.289.17136 wwwmatthes.in.tum.de