Web Science/Part3: Behavioral models/MoocIndex
Appearance
--MoocIndex for MOOC @ Web Science/Part3: Behavioral models
lesson|Processes on the Web Graph: The Example of Modelling the Dynamics of Meme Spreading
[edit | edit source]- video=File:Web science mooc flipped class room spreading memes.webm
- learningGoals=
- understand that network structure determines processes, such as individual communication
- understand that the network structure determines global communication results
- understand how to model micro-behavior of individuals at large
- understand how to related dying and exploding memes to the same model
- understand the difference of perspectives between micro interactions and macro effects
- Know http://www.nature.com/srep/2012/120329/srep00335/full/srep00335.html
- Know about effective distance http://link.springer.com/article/10.1140%2Fepjb%2Fe2011-20208-9 http://rocs.hu-berlin.de/D3/ebola/
- furtherReading=
unit|Overview of the phenomenon
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a
- b
- c
unit|Experimental Setup and Methodology of the Memes spreading Model
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a
- b
- c
unit|Mathematical foundations of the Memes spreading Model
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a
- b
- c
unit|Results of the Memes spreading Model
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a
- b
- c
unit|Summary, Further readings, Homework
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a
- b
- c
lesson|More Micro Behavior and Macro Effect I: Collective Intelligence
[edit | edit source]- learningGoals=
- Know about examples of collective intelligence in the Web (and beyond)
- Understand that clever aggregation of randomly noisy sensor output leads to high quality measurements
- Understand that independence of judgement is key to high quality collective decision making
- Relate this to law of large numbers
- Understand the idea of a social sensor: Model people output as sensor output
- Understand the idea of recursive aggregation of reputation
- Understand limitations of when collective intelligence cannot be derived
unit|IDF as Simple Form of Collective Intelligence
[edit | edit source]- learningGoals=
- IDF aggregates common usage of vocabulary
- knowledge about common usage of vocabulary models term specificity
unit|In-degree as Form of Collective Intelligence
[edit | edit source]- learningGoals=
- IDF aggregates common usage of vocabulary
- knowledge about common usage of vocabulary models term specificity
unit|Random surfer Model
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a
- b
- c
unit|Page rank of Graph/Matrix
[edit | edit source]- learningGoals=
- Eigenvalues are an important metric to describe graphs.
- Decomposing large matrices is computationally heavy.
- relation to the random surfer model
lesson|More Micro Behavior and Macro Effect II: Herding
[edit | edit source]- furtherReading=
- understand: https://www.princeton.edu/~mjs3/salganik_dodds_watts06_full.pdf
- know some basics about: Herd behavior from the field of psychology
- w:Absolute_difference
- w:Randomized_experiment
- w:Randomized_controlled_trial
- w:Randomization
- w:Web-based_experiments
- w:Conditional_independence
- w:Independence_(probability_theory)
- w:Dependent_and_independent_variables
- w:Herd_behavior
- w:Systematic_error
- learningGoals=
- Know and understand the notion of herding and swarms
- Know and understand that local information and positive feedback cycles may destroy collective intelligence (e.g. Groupthink, shitstorms, Klaas' tagging experiments, stock exchange.....)
- Know about examples of herding, such as preferential attachment, music experiment,...
- Understand how herding can be measured in an experiment
- How to conduct a web based experiment with a control group?
- Get to know one specific experiment and methodology that demonstrated herd behavior on the web.
- Understand how to empirically design an experiment that can demonstrate herd behavior.
- Discussing systematic errors in experiments
- Understand that it is non trivial to verify phenomenons of herding.
- understand: https://www.princeton.edu/~mjs3/salganik_dodds_watts06_full.pdf
- video=File:Web science mooc recommendations.webm
unit|Research question of herd behavior, inequality and unpredictability of cultural markets
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- What are the research questions that will be answered in the experiment
- understand that a good study starts with a research question
- The concept of falsifiability.
- Good research questions often start with an obervation (e.g.: experts have frequently failed to predict the success of musicians)
- furtherReading=
unit|Experimental Setup and data collection process
[edit | edit source]- learningGoals=
- difference between the dependent and independent group
- what is scientific control
- Repetition of the experiment (Why do the authors have 8 worlds?) to to conduct a randomized experiment.
- furtherReading=
- w:Dependent_and_independent_variables
- w:Independence_(probability_theory)
- w:Treatment_and_control_groups
- w:Randomized_experiment
- w:Randomized_controlled_trial
unit|Discussion of Systematic errors
[edit | edit source]- learningGoals=
- Critical discussion of the web limitations that are posed in the paper. (web scientists can get rid of some of these mistakes)
- Understand that systematic errors are part of many experiments.
- Learn to discuss systematic errors of a paper.
- which measures have been taken to minimize the amount of systematic errors (e.g. introducing 8 worlds)
- furtherReading=
unit|Metrics and their mapping to the research questions
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a measure for inqueality: the gini coefficient
- unpredictability needs the 8 worlds to see how different rankings are
- market share
- furtherReading=
- w:Gini_coefficient
- w:Mean_difference for unpredictability
unit|Results of the Music Recommendation hearding experiments
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- we can observe clear hearding behavior.
- the way conent is presented on the web has an impact of how people consume it.
- c
unit|Summary, Further readings, Homework
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- music experiments are just one empirical indicator for hearding behavior
- other behavior might night another scientific methodology to identify the behavior.
- Dellschaft shows that herding may reduce quality of information categorization ([1])
- More on herding: link to http://slon.ru/upload/iblock/4a1/Science-2013.pdf
lesson|User modelling, personalizing, collaborative filtering, Recommendations
[edit | edit source]- learningGoals=
- don't know where to place this lesson yet. It should somehow point out how collective intelligence is used for recommendations and how this is influenced
- furtherReading=
unit|user modelling
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- furtherReading=
unit|Collaborative filtering
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- furtherReading=
unit|Recommendation
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- furtherReading=
unit|Personalizing content
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- furtherReading=
unit|Summary, Further readings, Homework
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- furtherReading=
lesson|Advertisement Ecosystems
[edit | edit source]- learningGoals=
- Understand how cross-site advertisement providers function on the Web
- Understand advertisement KPIs
- Relate to recommendations
unit|Introduction to Online Advertisement
[edit | edit source]- furtherReading=
- w:Online_advertising
- http://www.rene-pickhardt.de/retargeting-smart-online-marketing-system-by-criteo/
- http://www.iab.net/media/file/IAB_Internet_Advertising_Revenue_Report_FY_2013.pdf and http://www.iab.net/research/industry_data_and_landscape/adrevenuereport
- learningGoals=
- understand the interests of the 4 players (publisher (content owner), advertiser (some brand), ad-service, consumer)
- be aware of the online ad market and be able to relate it to other ad markets
- be aware of advertising formats
- be aware of payment formats for online advertisement
- test edit
- video=File:Introduction_to_Online_Advertisement.webm
unit|Metrics for (online) advertisement
[edit | edit source]- furtherReading=
- http://tlvmedia.com/pdf/CPM_CPC_CPA_dCPM.pdf
- w:Cost_per_mille
- w:Click-through_rate
- w:Pay_per_click
- w:Affiliate_marketing and w:Cost_per_acquisition
- w:Bounce_rate
- w:Conversion_rate
- learningGoals=
- be able to list basic metrics of online advertisement (CPC, CTR, CR, BR, CPM) and calculate them
- be able to interpret the metrics.
- understand which player should optimize which metric
- video=File:Metrics_for_online_advertisement.webm
unit|Factors that have impact on advertisement campaigns
[edit | edit source]- furtherReading=
- w:Conversion_optimization
- w:Landing_page_optimization
- w:Bait-and-switch
- w:Frequency_capping
- w:Lead_scoring
- w:Targeted_advertising
- w:Negative_keyword (very interesting, it shows the amount of data Google has due to ad products)
- w:Online_advertising#Trick_banners
- w:Behavioral_targeting
- w:Contextual_advertising
- learningGoals=
- Relevance
- Targeting (which is a form of relevance)
- User Context
- Truthfulness of the add
- design of the landing page (usability)
- test
- video=File:Factors_impact_on_advertisement_campaigns.webm
unit|Finding the true value of an advertisement
[edit | edit source]- furtherReading=
- w:Second_price_auction
- w:Auction_theory
- w:Game_theory
- w:Nash_equilibrium
- w:Generalized_second-price_auction
- original literature: paper and slides
- learningGoals=
- Second price auctions
- Collective intelligence
- video=File:Under_construction_icon-blue.svg
unit|Understanding the Problems with Click Fraud
[edit | edit source]- furtherReading=
- learningGoals=
- Understand reasons why people would produce click fraud
- Understand to whom click fraud is harmful.
- video=File:Understanding_problems_with_click_fraud.webm
unit|Summary, Further readings, Homework
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a
- b
- c
lesson|social capital and rational choice theory
[edit | edit source]unit|unit 1
[edit | edit source]- learningGoals=
- furtherReading=
unit|unit 2
[edit | edit source]- learningGoals=
- furtherReading=
unit|unit 3
[edit | edit source]- learningGoals=
- furtherReading=
unit|unit 4
[edit | edit source]- learningGoals=
- furtherReading=
unit|Summary, Further readings, Homework
[edit | edit source]- video=File:Under_construction_icon-blue.svg
- learningGoals=
- a
- b
- c