Data Science and Artificial Intelligence Boot camp

✓ Salary Hike: 40% to 80% higher than regular Software Engineer.

✓ Growth Rate: 290% increase in Machine Learning – Artificial Intelligence Jobs in India

✓ 100% Practical Sessions

✓ Live Projects to Work

✓ Certificate of recommendation after successful projects

Introduction
1. Understanding the core idea of machine learning a computer to learn concepts
using data.
2. Different use cases
3. Limitation and scope
1 class of 1 hour

Basic Mathematics / ( mainly Linear Algebra )
The basic mathematics required for the course.
Assignments : 2
Quiz : 1
2/3/4 classes : 1 Hour each

Hands on Python
Introduction to python
Basic code syntax
Installation and setup of python ( Jupyter Notebook )
Understanding the libraries required for ML
Assignments : 4 ( each can be solved in 30 minutes approx )
Quiz : 1

Depending on requirement can vary from 7 – 15 classes : 1 Hour each
My basic approach is to introduce one ML problem before the module two. Which can
be used as baseline for understanding next concepts requires one class

Types of Machine Learning And Performance Measurement
( Mainly theoretical )
1. Discussion about Supervised and unsupervised ML
2. Regression and Classification Models
3. How and when to choose which models
4. Measuring Performance in Regression Models
a. Quantitative Measures of Performance
(Includes accuracy measures , F1 Scores , Recall , etc)
5. Measuring Performance in Classification Models
a. Class Predictions
b. Evaluating Predicted Classes
c. Evaluating Class Probabilities
6. The Variance-Bias Trade-off
7. Different types of Loss Function

Quiz
Assignment : Introduction to a real world ML problem
Requires 7-10 hours Important one
Note – A lot of topics and questions are related to this topic some are not
mentioned but if someone asks can be added.
This will give an overall understanding and knowledge of ML.
One/two basic algorithms can be introduced.

Data Preprocessing
Understanding Importance of Data Pre-processing
It includes
1. case study
2. Transformations for Individual Predictors
3. Data Transformations for Multiple Predictors
4. Dealing with Missing Values
5. Removing Predictors
6. Adding Predictors
7. Binning Predictors
2 Assignment ( requires Python )
3/4 classes 1 hour each

Over-Fitting and Model Tuning
The Problem of Overfitting
1. Model Tuning
2. Data Splitting
3. Resampling Techniques
4. Choosing Final Tuning Parameters
5. Data Splitting Recommendations

2 Quiz
2 classes 1 hour each

ML ALGORITHMS
Supervised Learning
Linear Regression ( Single and Multivalued)
1. Linear Regression
2. Partial Least Squares
3. Penalized Models
Nonlinear Models
1. Multivariate Adaptive Regression Splines
2. Support Vector Machines
3. K-Nearest Neighbors
4. Neural Networks
Trees and Rule-Based Models
1. Basic Regression/Classification Trees
2. Regression/Classification Model Trees
3. Rule-Based Models
4. Bagged Trees
5. Random Forests
6. Boosting

Bayesian Models
1. Naive Bayes
Unsupervised Learning
1. Dimensionality Reduction
a) Principal Components Analysis
b) TF-IDF
2. Clustering Algorithms
For each of the above mentioned algorithms following this will be covered :
1) Mathematical derivation
2) Associated preprocessing that can be useful
3) Associated Loss functions
4) Scenarios where they can be suited best
5) Limitations
We will solve around 4 ML based problems ( depending upon the time constraint ).
The problems will vary from basic to industrial level.
Assignments and quiz for each algorithm.

Timeline for module 4
1. Each algorithm require around 1 hour for introduction derivation .
2. In some cases some external concepts need to be taught like in tree
based models so they require 2 hours .
3. Depending upon the timeline no of algorithms can be increased/
decreased .

Deep Learning/AI Syllabus:

1. Foundations of Neural Networks :

1. Introduction to deep learning
2. Neural networks basics
3. Shallow neural networks
4. Deep neural networks

2. Structuring Machine Learning Projects

1. ML Strategy, Setting up your goal, human level performance.
2. Error Analysis, Mismatched training and dev/ test distributions, learning for multiple tasks, end-to-end deep learning.

 

3. Convolutional models

1. Convolutional neural networks
2. Computer vision applications

4. Sequence models

I RNN, LSTM, GRU models II Application to NLP

5. Case studies (2 )

I In-depth discussion of the prpb lem statement

II Discussion of student projects.