The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio.
Challenges for IoT in Industrial Automation Lifecycle (>15 years)
Robust, highly available
Well supported
Closed
Diversity
Incremental changes
Small budgets
High data intensity
Security
IoT trackrecord (“we don’t want our competitor to know”)
USP IoT (“we already have that”)
Maintenance staff
Our technology has gotten smart and fast enough to make predictions and come up with recommendations in near real time. Machine Learning is the art of deriving models from our Big Data collections – harvesting historic patterns and trends – and applying those models to new data in order to rapidly and adequately respond to that data. This presentation will explain and demonstrate in simple, straightforward terms and using easy to understand practical examples what Machine Learning really is and how it can be useful in our world of applications, integrations and databases. Hadoop and Spark, real time and streaming analytics, Watson and Cloud Datalab, Jupyter Notebooks, Oracle Machine Learning CS and the Citizen Data Scientists will all make their appearance, as will SQL.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio. This presentation discusses architecture trends, container technology, disruptive movements such as IoT, Blockchain, Intelligent Bots and Machine Learning, Modern User Experience, Enterprise Integration, Autonomous Systems in general and Autonomous Database in particular, Security, Cloud, Networking, Java, High PaaS & Low PaaS, DevOps, Microservices, Hybrid Cloud. This Oracle OpenWorld - more than any in recent history - rocked the foundations of the Oracle platform and opened up some real new roads ahead. This presentation leads you through the most relevant announcements and new directions.
Bridging the gap between Administrative and Operational IT
Vision, Architecure and Project experience. This slide deck shows our vision on this market for industrial enterprise IOT. Conclusion
Challenges for IoT in Industrial Automation Lifecycle (>15 years)
Robust, highly available
Well supported
Closed
Diversity
Incremental changes
Small budgets
High data intensity
Security
IoT trackrecord (“we don’t want our competitor to know”)
USP IoT (“we already have that”)
Maintenance staff
Our technology has gotten smart and fast enough to make predictions and come up with recommendations in near real time. Machine Learning is the art of deriving models from our Big Data collections – harvesting historic patterns and trends – and applying those models to new data in order to rapidly and adequately respond to that data. This presentation will explain and demonstrate in simple, straightforward terms and using easy to understand practical examples what Machine Learning really is and how it can be useful in our world of applications, integrations and databases. Hadoop and Spark, real time and streaming analytics, Watson and Cloud Datalab, Jupyter Notebooks, Oracle Machine Learning CS and the Citizen Data Scientists will all make their appearance, as will SQL.
The annual review session by the AMIS team on their findings, interpretations and opinions regarding news, trends, announcements and roadmaps around Oracle's product portfolio. This presentation discusses architecture trends, container technology, disruptive movements such as IoT, Blockchain, Intelligent Bots and Machine Learning, Modern User Experience, Enterprise Integration, Autonomous Systems in general and Autonomous Database in particular, Security, Cloud, Networking, Java, High PaaS & Low PaaS, DevOps, Microservices, Hybrid Cloud. This Oracle OpenWorld - more than any in recent history - rocked the foundations of the Oracle platform and opened up some real new roads ahead. This presentation leads you through the most relevant announcements and new directions.
Bridging the gap between Administrative and Operational IT
Vision, Architecure and Project experience. This slide deck shows our vision on this market for industrial enterprise IOT. Conclusion
2. Inhoud
• Introductie in R
• R syntax
• Introductie Decision Tree model
• Voorbeeld Decision Tree model
• Handson met Decision Trees
Introductie in R 2
4. Introductie R
• High level scripttaal
• Ontwikkeld voor statistische
berekeningen
• Visualiseert modellen zeer gemakkelijk
• Vergelijkbaar met Matlab en Python
Introductie in R 4
5. Introductie R
• Grootste marktaandeel binnen data
science
• Gratis en open source
• Grote hoeveelheid libraries
• Grote community
Introductie in R 5
6. Installatie R
• Download en installeer R:
• https://www.r-project.org/
• Download en installeer Rstudio (een
goede en gratis IDE voor R)
• https://www.rstudio.com/
Introductie in R 6
7. Introductie in R 7
R syntax
Toekenning
• Merk op: de x’te machtswortel van y is y^(1/x)
Berekeningen
8. Introductie in R 8
R syntax
• Merk op: vectoren beginnen bij 1!
• [0] geeft het type vector
• Out of bounds geeft NA, geen error
Vectoren
• Merk op: default operaties zijn scalair
• Matrices beginnen bij 1.
Matrices
9. R syntax
• Packages installeren en aanroepen:
(Je “import” statements)
Introductie in R 9
11. Intuïtief voorbeeld
Je vraagt aan een vriend: “Welke sport
moet ik gaan doen?”
• Teamsport?
• Hou je van balsporten?
• Werk je het liefst met je handen?
• Hou je van stoeien?
• Doe je graag zwaar werk?
Introductie in R 11
Teamsport
Balsport
Handen
Basketbal Voetbal Stoeien
ZwaarJiu jitsu
Krachttraining Hardlopen
Touwtrekken
12. Introductie
• Classification and Regression Tree
(CART)
• Zeer gemakkelijk qua interpretatie
• Predictief niet het beste
Methoden om CART te verbeteren:
• Bagging
• Boosting
• Random Forest
• Voorbeeld hiernaast: huisprijs
Introductie in R 12
13. Voordelen
• Conceptueel gemakkelijk
• Gemakkelijke interpretatie
• Werkt even goed voor classificatie als
regressie
• Snel
• Gaat goed om met missing values
Introductie in R 13
15. Opbouw
• 1 grote dataset: veel
onzekerheid/variantie
• Opsplitsen in categorieën, steeds
minder variantie per categorie
• Opsplitsing die de meeste variantie
verklaart eerst
• Dan de volgende binnen die
categorie
Introductie in R 15
16. Interpretatie
• Voldoet aan conditie: links, anders:
rechts
• Bovenaan: belangrijkste opsplitsing
• Lengte van de “tak” geeft aan hoe
veel variantie verklaard wordt door
deze variabele
Introductie in R 16
17. Titel van de presentatie 17
Voorbeeld: huisprijs
Woningprijzen inladen
• Prijs afhankelijk van periode en regio
• Summary geeft een overzicht van hoe goed de
voorspelling is (residuals)
Decision tree bouwen
• Package installeren en laden
• Working directory setten
• CSV inladen
• Head() toont de eerste regels van de dataset
18. Data Science 101
1. Inspecteer de dataset
2. Denk na: wat zit er in de data
3. Stel dan een model op dat voor jou
logisch is. Wat is je afhankelijke? Wat
zijn je onafhankelijken?
Je moet kunnen verklaren waarom je
dat model hebt opgesteld!
4. Bouw je model in R
5. Interpreteer je model
Introductie in R 18
19. Titel van de presentatie 19
Hands on met Decision Trees!
Moeilijk:
https://github.com/AMIS-Services/machine-learning-
session-one-7-may-2018/tree/master/technical-handson
Makkelijk:
https://bit.ly/2HVgKat
Tips:
Inspecteer eerst de data!
Stel dan een model op dat voor jou logisch is.
Maak je model in R.
Interpreteer je model.
20. Naslagwerk
Voor zij die meer willen weten:
http://www-
bcf.usc.edu/~gareth/ISL/ISLR%20First%20
Printing.pdf
Een uitgebreid naslagwerk over
gebruikelijke machine learning methoden en
de achterliggende theorie.
Aan te raden als je dieper in wilt gaan op de
achterliggende wiskunde en theorie.
Introductie in R 20