Professional Documents
Culture Documents
Overview of course:
Management as decision orient activity
How to make decision using more data and knowledge (getting rigth info for right decision)
Consequence of digitilzation, societym orgiznation and individual.
Future role: business translator, consultant company, n
Need to have deeper knowledge and industry or functional expertise
Management is science or art about getting thing done. (in contect of Digitilization)
Mobilization of Collective action with other (work together)
System 1 quick decision makign based on experiecne logic with blink gut feelings, judgements
System 2 more scienctific approach, , matrix. Analytical models
1. Data
2. Descriptive (experiecene jsudgements)
3. Predictive (use data to predict future, regression coorelation)
4. Prescriptive (use data to give order)
Analystical models
1. Framing assumption
2. Naming
3. Relating
Lecture 2:
Lecture 3:
Chapter 1 Data Science by Kelleher
Date Science encompasses a set of prinicnples problem definitions algorithems and pricess for
extracting non obvious and useful patterns from large data sets.
Date scienec machine learning (design and evaluation of algorithms) and data mining (analysis
of structured data) are interchangeable focus on improving decsision making throuhj the analysis
of data.
Clustring/Customer Segmentation: identifying group of customers exhibiting similar behavior
and teastes.
Association rule minning: indentiies products that are frequently bought together
Anomaly/ Outlier detection: identifying strange or abnormal events (fraudulent clains)
Prediction: identifying these types of classification patterns.
Actionable insight: using the context to describe what we want the extracted patterns to give
us.
Transactional Data vs non transactional data:
Relational Data Model how data were stored indexed and retrived from database.
Meta data: data describing the structure and properties of the raw data.
Big Data defined in terms of: Volume variety and velocity.
Mapreduced framework: where data is distributed on multiple server and partial results
calculated on each servers are merged/ reduced.
Data analysis: descriptive (summerization) vs statistics (measurement)
All of the data in these various databases and data sources will need to be integrated, cleansed,
transformed, normalized, and so on. These tasks go by many names, such as extraction,
transformation, and load , “data munging,” “data wrangling,” “data fusion,” “data crunching,
Data architecture:
Describe how data is collected stored transformed distributed and consumed.
Lecture 4:
Mind- a theroy of mind: without understading of our thought and behavior.
Cognition: how we think (process) and feel (emotion).
Utility theory: Explain behavior of individuals based in their choices/ preferecnes. (governence
of pain and pleasure)
Rational choice theory: cost benefit analysis to make choices
Prospect theory: diffrenet situation due to different defination/naming of the siutation.
Discriptive statistics vs probablilities *deals with unknown
Rational choice: expected value (the bell curve, statistics), expected utility
S curve: /Marginal diminishing law
Framing: reference point/
Risk avers/ risk seeking
Framing and Preference:
In many situations we don’t know what we prefer. It changes with situation.
1. Interconnected decision:
2. Pseudocartainty: ppl value the creatoin of certainty over and equally valed shift
3. Whats is worth to you:
4. Endownment efftect: value we place on what we own
5. Mental accounting: different decsion rules to different mental accounts
Heuristics:
Ppl rely in number of sumplifying strategies/rule of thumb when making decisions.(learning and
making deicision with personal understing) self discouvery. Time pressured and time bounded,
experiments.
Types of heuristic:
1. Availability of heuristic: we take what we have/readily avaiable in memory
Leads to cognitive biases (ease of recall and retrievability)
Complexity:
1. Difficult to identify casue and effect, causality
2. Systems- intergrated interdependents modules
3. Power laws small casue large consequences
Machine learning:
Restrictions
1. Cant codify formal rules
2. High cost
3. Computer power (quantum computing)
Symbolists vs comnnectionism:
Machine learning can be
1. Supervised: classficication and regression (decision tree, neutral networks)
2. Unsupervised: clusting
Deep learning:learns hierarchicals represtation from the data itself and scales with more data.
Black box: accountability(provision of reasions, explanations and justifications)
Epistemic (what kind of data) normative standards (how we use it)
Limoncelli: 10 things Executives should know about software
1. software is not magic
2. Software is never done
3. Software is team effort
4. Design isnt how something looks. It is how it works.
5. Security is everything (cybersecutiy, risk)
6. Feature size doesn’t predict developer time
7. Greatness comes from thousands of small improvements
8. Technical debt is bad but unavoidable
9. Software doesn’t run itself
10. Complex system need devops to run well
Process redesign.
Roles; new expectations relations responsibitlies create demands machine capable roels.
Conflicting goals- need of balance
1. Different goals (shareholder&stakeholder
2. Grwoth & profitiablity
3. Efficiency *stability) and Effectiveness flexible adabptibility
4. Time frames – short or long term
Analytical approach:
1. Information processing
2. Interpertive approach
3. Politacl approach ( how different interest
Univerisal problem: task division, task allocation, provision of reqards and provision of
information, exception management.
1: RESPECTFUL INQUIRY
Self determination theory (that all humans have basic psychological needs, which social
interactions)
Under Behavious, help setting goals feedback, decision making. Integrated leadershp field.
Governance Prespective:
Commons principles: Conflict, mutuality and order
Governence structure incentive intensity, administrative control and contact law regim.
Behavioral assumtions:
1. Bounded Rationality
2. Opportunism
Computational approaches:
1. Differential privacy
2. Federated learning