Hi.. Hello and welcome to my new course, Machine Learning with Python for Dummies. We will discuss about the overview of the course and the contents included in this course.

Artificial Intelligence, Machine Learning  and Deep Learning Neural Networks are the most used terms now a days in the technology world. Its also the most mis-understood and confused terms too.

Artificial Intelligence is a broad spectrum of science which tries to make machines intelligent like humans. Machine Learning and Neural Networks are two subsets that comes under this vast machine learning platform

Lets check what's machine learning now. Just like we human babies, we were actually in our learning phase then. We learned how to crawl, stand, walk, then speak words, then make simple sentences.. We learned from our experiences. We had many trials and errors before we learned how to walk and talk. The best trials for walking and talking which gave positive results were kept in our memory and made use later. This process is highly compared to a Machine Learning Mechanism

Then we grew young and started thinking logically about many things, had emotional feelings, etc. We kept on thinking and found solutions to problems in our daily life. That's what the Deep Learning Neural Network Scientists are trying to achieve. A thinking machine.

But in this course we are focusing mainly in Machine Learning. Throughout this course, we are preparing our machine to make it ready for a prediction test. Its Just like how you prepare for your Mathematics Test in school or college.  We learn and train ourselves by solving the most possible number of similar mathematical problems. Lets call these sample data of similar problems and their solutions as the 'Training Input' and 'Training Output' Respectively. And then the day comes when we have the actual test. We will be given new set of problems to solve, but very similar to the problems we learned, and based on the previous practice and learning experiences, we have to solve them. We can call those problems as 'Testing Input' and our answers as 'Predicted Output'. Later, our professor will evaluate these answers and compare it with its actual answers, we call the actual answers as 'Test Output'. Then a mark will be given on basis of the correct answers. We call this mark as our 'Accuracy'. The life of a machine learning engineer and a data-scientist is dedicated to make this accuracy as good as possible through different techniques and evaluation measures.

Here are the major topics that are included in this course. We are using Python as our programming language. Python is a great tool for the development of programs which perform data analysis and prediction. It has tons of classes and features which perform the complex mathematical analysis and give solutions in simple one or two lines of code so that we don't have to be a statistic genius or mathematical Nerd to learn data science and machine learning. Python really makes things easy.

These are the main topics that are included in our course

System and Environment preparation

-----------------------------------

Installing Python and Required Libraries (Anaconda)

Basics of python and sci-py

---------------------------

Python, Numpy , Matplotlib and Pandas Quick Courses

Load data set from csv / url

-----------------------------

Load CSV data with Python, NumPY and Pandas

Summarize data with description

--------------------------------

Peeking data, Data Dimensions, Data Types, Statistics, Class Distribution, Attribute Correlations, Univariate Skew

Summarize data with visualization

-----------------------------------

Univariate, Multivariate Plots

Prepare data

-------------

Data Transforms, Rescaling, Standardizing, Normalizing and Binarization

Feature selection – Automatic selection techniques

-----------------------------------

Univariate Selection, Recursive Feature Elimination, Principle Component Analysis and Feature Importance

Machine Learning Algorithm Evaluation

-----------------------------------

Train and Test Sets, K-fold Cross Validation, Leave One Out Cross Validation, Repeated Random Test-Train Splits.

Algorithm Evaluation Metrics

-----------------------------

Classification Metrics - Classification Accuracy, Logarithmic Loss, Area Under ROC Curve, Confusion Matrix, Classification Report.

Regression Metrics - Mean Absolute Error, Mean Squared Error, R 2.

Spot-Checking Classification Algorithms

-----------------------------------

Linear Algorithms -  Logistic Regression, Linear Discriminant Analysis.

Non-Linear Algorithms - k-Nearest Neighbours, Naive Bayes, Classification and Regression Trees, Support Vector Machines.

Spot-Checking Regression Algorithms

-----------------------------------

Linear Algorithms -   Linear Regression, Ridge Regression, LASSO Linear Regression and Elastic Net Regression.

Non-Linear Algorithms - k-Nearest Neighbours, Classification and Regression Trees, Support Vector Machines.

Choose The Best Machine Learning Model

-----------------------------------

Compare Logistic Regression, Linear Discriminant Analysis, k-Nearest Neighbours, Classification and Regression Trees, Naive Bayes, Support Vector Machines.

Automate and Combine Workflows with Pipeline

-----------------------------------

Data Preparation and Modelling Pipeline

Feature Extraction and Modelling Pipeline

Performance Improvement with Ensembles

-----------------------------------

Voting Ensemble

Bagging: Bagged Decision Trees, Random Forest, Extra Trees

Boosting: AdaBoost, Gradient Boosting

Performance Improvement with Algorithm Parameter Tuning

--------------------------------------------------------

Grid Search Parameter

Random Search Parameter Tuning

Save and Load (serialize and deserialize) Machine Learning Models

-----------------------------------

Using pickle

Using Joblib

finalize a machine learning project

-----------------------------------

steps For Finalizing classification models - pima indian dataset

Dealing with imbalanced class problem

steps For Finalizing multi class models - iris flower dataset

steps For Finalizing regression models - boston housing dataset

Predictions and Case Studies

----------------------------

Case study 1: predictions using the Pima Indian Diabetes Dataset

Case study: Iris Flower Multi Class Dataset

Case study 2: the Boston Housing cost Dataset

Machine Learning and Data Science is the most lucrative job in the technology arena now a days. Learning this course will make you equipped to compete in this area.

Best wishes with your learning. Se you soon in the class room.