Certified Machine Learning Specialist (CMLS)
description
Transcript of Certified Machine Learning Specialist (CMLS)
Certified Machine Learning Specialist
(CMLS)
Course Outline
www.globalicttraining.com
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.
Course Outline
Machine Learning is the science of getting computing systems to act with minimal human
Intervention. Such systems change behaviors without being explicitly programmed. It is
nothing but automation of Feature Engineering
With Machine Learning factors such as increasing volume and variety of available data,
cheaper and powerful computational processing and affordable data storage will result in
high-value predictions that can guide better decisions.
Gartner predicts that by 2020, Artificial Intelligence will be one among the top five investment
priority for more than 30 percent of CIOs. Machine Learning is going to be a critical driver for
demand, fraud and failure predictions by 2019.
This specialist course on Machine Learning gives an overview of many concepts, techniques,
and algorithms. The course covers basic topics such as classification and linear regression to
more advanced topics such as boosting, support vector machines, hidden Markov models, and
Bayesian networks in Machine Learning.
This course will introduce professionals to open source Machine Learning tools such as WEKA
and Scikit Learn. Professionals attending the course will learn the different algorithms in some
of most widely adopted Machine Learning methods such as Supervised Learning,
Unsupervised learning and Reinforcement Learning.
JOB ROLES IN NICF / TARGETED AUDIENCE
• Data Analyst - Statistics and Mining
• Data Analyst - Text Analytics
• Operations Research Analyst
• Senior Data Analyst- Statistics and Mining
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.
PRE-REQUISITES
Participants are preferred to have experience in software development, business domain or
data/business analysis.
PROGRAM STRUCTURE
This is a 4-day intensive training program with the following assessment components.
Component 1. Written Examination
Component 2. Project Work Component
These components are individual based. Participants will need to obtain 70% in both the
components in order to qualify for this certification. If the participant fails one of the
components, they will not pass the course and have to re-take that particular failed
component. If they fail both components, they will have to re-take the assessment.
Course Outcomes
• Learn the fundamentals of AI and machine learning and how it could impact your workthrough several real-life use case
• Understand Machine Learning techniques/method: Supervised, Unsupervised &Reinforcement Learning through hands-on examples
• Learn key ML concepts like Principle Component Analysis (PCA), Hyperparameter tuning,Clustering, Classification ,Regression ,Neural Network etc.
• Get skilled in popular machine learning algorithms using Python Programming( SckitLearn, TensorFlow ) ,Weka, RapidMiner
•
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.
Course Structure
Day Session 1
(9:00 – 10:45)
Session 2
(11:00 – 12:30)
Session 3
(13:45 – 15:15) Session 4
(15.15 – 4.45)
Day 1 Introduction and Basic concepts in Machine learning
Introduction to Theories used in
Machine Learning
Supervised learning vs.
Unsupervised learning
Model Selection in Machine learning
Day 2 Role of Weka in Machine Learning
Decision Tree and Rule Mining
using Weka
Hands-On (WEKA)
Hands-On (WEKA)
Day 3 A Brief review on
SciPy
Random Forest and Markov
Decision Process Algorithm
Hands-On
(SciKit
learn)
Hands-On
(Scikit
Learn)
Day 4
Google’s Go Programming with k-nearest
Neighbor’s Algorithm
Hands-On (Spark MLLIB)
Review and Discussion
Written Examination
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.
COURSE OUTLINE
Unit 1: Introduction and basic concepts in Machine learning
- Definition of machine learning systems
- Goals and applications of machine learning
- Classification of machine learning algorithms
- Supervised learning
- Unsupervised learning
- Reinforcement learning
Unit 2: Introduction to Theories used in Machine Learning
- Introduction to probability theory
- Discrete random variables and Fundamental rules
- Independence and conditional independence
- What is Information theory?
- How Decision Theory is helpful in machine learning
- Learning Theory of the machine learnings
Unit 3: Supervised learning vs. Unsupervised learning
- Supervised learning setup
- Logistic regression
- Gaussian discriminant analysis and Naive Bayes
- Support vector machines
- Clustering and K-means
- PCA (Principal components analysis)
- ICA (Independent components analysis)
- Evaluating and debugging learning algorithms
Unit 4: Model selection in Machine learning
- Bayesian model selection
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.
- Model selection for probabilistic models
- Model selection for non-probabilistic methods
- Probabilistic Generative Models
- Probabilistic Discriminative Models
Unit 5: Role of Weka in Machine Learning
- Introduction to weka
- How to install Weka
- The Knowledge Flow interface
- The Command Line interface
- Classification Rules and association Rules
- Attribute Selection and Fast attribute selection using ranking
Unit 6: Decision Tree and Rule mining using Weka
- ID3 based decision tree algorithm
- Entropy and Information gain
- ID3 implementation using weka
- Association rule mining using Frequent Pattern (FP) Growth algorithm
- FP-Tree structure
- FP-Growth Algorithm
- Implementation of FP-Growth using weka
Unit 7: A Brief review on SciPy
- SciPy - A introduction
- Scipy installation in ubuntu
- Python Scientific Computing Environment
- The SciPy Library/Package
- Data structures and function of Scipy
- Numpy
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.
Unit 8: Random Forest and Markov Decision Process algorithm
- Decision tree learning
- Tree bagging
- Random forests generation
- Relationship to nearest neighbors
- Markov model and Hidden Markov Model (HMM)
- Inference in HMM
- Learning and generalizations of HMM
Unit 9: Google’s Go Programming with k-nearest neighbor’s algorithm
- Introduction to Go Programming
- Mark Bates on Go Core Techniques and Tools
- Mark Bates on Go Database Web Frameworks and Techniques
- k-nearest neighbor’s algorithm
- Parameter selection
- Metric learning and Feature extraction
- Dimension reduction and Decision boundary
- Selection of class-outliers
Unit 10: C 5.0 based decision tree algorithm
- Introduction to C4.5 and C 5 Decision Tree
- Divide and Conquer Technique
- Feature Selection
- Regression Trees
- Selecting and Candidate testing
- Estimating True Error rate
- Pruning Decision Trees
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.
Hands On
The main objective of this course is to train the professional with open source tools such as
Google GO, Scikit Learn and WEKA, Spark MLLIB.
WRITTEN ASSESSMENT
As part of the written examination, each participant will be assessed individually on the last
day of the training for their understanding of the subject matter and ability to evaluate, choose and
apply them in specific context and also the ability to identify and manage risks. The assessment
focuses on higher levels of learning in Bloom’s taxonomy: Application, Analysis, Synthesis and
Evaluation. This written examination will primarily consist of 40 multiple choice questions spanning
various aspects as covered in the program. It is an individual, competency-based assessment.
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.
EXAM PREPARATION
The objective of the certification examination is to evaluate the knowledge and skills acquired by the
participants during the course. The weightage in key topics of the course as follows:
• Introduction and basic concepts in Machine learning (10%)
• Introduction to Theories used in Machine Learning (10%)
• Supervised learning vs. Unsupervised learning (10%)
• Model selection in Machine learning (10%)
• Role of Weka in Machine Learning (10%)
• Decision Tree and Rule mining using Weka (10%)
• A Brief review on SciPy(10%)
• Random Forest and Markov Decision Process algorithm (10%)
• Google’s Go Programming with k-nearest neighbor’s algorithm and Spark MLLIB (10%)
• C 5.0 based decision tree algorithm (10%)
Tools Used:
• WEKA
• SCIPY
• Spark MLLIB
• GO
COPYRIGHT © GICT TRAINING & CERTIFICATION. ALL RIGHTS RESERVED.