Duration: Hours

MATLAB makes machine learning easy. With tools and also functions for handling big data, as well as apps to make machine learning accessible it is an ideal environment for applying machine learning to your data analytics.

Training Mode: Online

Enquiry

    Category: Tags: , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,

    Description

    Objectives

     a). How to implement different machine learning classification algorithms using matlab.

    b). How to impplement different machine learning clustering algorithms using matlab.

    c). How to proprocess data before analysis.

    d). When and how to use dimensionality reduction.

    e). Take away code templates.

    f). Visualization results of algorithms

    g). Decide which algorithm to choose for your dataset

     

    1. Introduction to course and MATLAB

    a). Course Introduction

    b). MATLAB essentials for the course

    2. Data Preprocessing

    a). Code and Data

    b). Section Introduction

    c). Importing the Datasets

    d). Removing Missing Data

    e). Feature Scaling

    f). Handling Outliers

    g). Dealing with Categorical Data

    h). Your Preprocessing Template

    3. Classifications

    a). Code and Data

    4. K – nearest Neighbour

    a). KNN Intuition

    b). KNN in MATLAB

    c). Visualizing the Decision Boundaries of KNN

    d). Explaining the code for visualization

    e). Here is our classification template

    f). How to change default options and customize classifiers

    g). Customization options for KNN

    5. Naive Bayes

    a). Naive Bayesian Intuition.

    b). Naive Bayesain in MATLAB

    c). Customization Options for Naive Bayesain

    6. Decision Trees

    a). Decision trees intuition

    b).Decision Trees in MATLAB

    c).Visualizing Decision Trees using the View Function

    d). Customization Options for Decision Trees

    7. Support Vector Machines

    a). SVM Intuition

    b). Kernel SVM Intuition

    c). SVM in MATLAB

    d). Customization Options for SVM

    8. Discriminant Analysis

    a). Discriminant Analysis Intuition

    b). Discriminant Analysis in MATLAB

    c). Customization Options for Discriminant Analysis

    9. Ensembles

    a). Ensembles Intuition

    b). Ensembles in MATLAB

    c). Customization Options for Ensembles

    10.Performance Evaluation

    a). Evaluating Classifiers: Confusion matrix (Theory)

    b). Validation Methods (Theory)

    c). Validation methods in MATLAB

    d). Evaluating Classifiers in MATLAB

    11. Clustering

    a). Code and Data

    12. K – means

    a). K-Means Clustering Intuition

    b). Choosing the number of clusters

    c). k-means in MATLAB

    d). KMeans Limitations

    13. Mean Shift Clustering

    a). Intuition of Mean Shift

    b). Mean Shift in MATLAB

    c). Mean Shift Performance in Cases where Kmean Fails

    14. DBSCAN

    a). Intuition of DBSCAN

    b). DBSCAN in MATLAB

    c). DBSCAN on clusters with varying sizes

    d). DBSCAN on clusters with different shapes and densities

    e). DBSCAN for handling noise

    15. Hierarchical Clustering

    a). Hierarchical Clustering Intuition

    b). Hierachical Clustering in MATLAB

    16. Dimensionality Reduction

    a). Code and data

    b). Principal Component Analysis

    c). PCA in MATLAB

    17. Data Preprocessing

    a). Code and data

    18. Handling Missing Values

    a). Deletion strategies

    b). Using mean and mode

    c). Considering as a special value

    d). Class specific mean and mode

    e). Random Value Imputation

    19. Dealing with Categorical Variables

    a). Categorical data with no order

    b). Categorical data with order

    c). Frequency based encoding

    d). Target based encoding

    20. Outlier Detection

    a). 3 sigma rule with deletion strategy

    b). 3 sigma rule with filling strategy

    c). Box plots and iterquartile rule

    d). Class specific box plots

    e). Histograms for outliers

    f). Local Outlier Factor

    g). Outliers in Categorical Variables

    21. Feature Scaling and Data Discretization

    a). Feature Scalling

    b). Discretization using Equal width binning

    c). Discretization using Equal Frequency binning

    Requirements

    a). MATLAB 2017a or heigher version. No prior knowledge of MATLAB is required

    b). In version below 2017a there might be some functions that will not work

    Basic Course Description 

    This course is for you if you want to have a real feel of the Machine Learning techniques without having to learn all the complicated maths. Additionally, this course is also for you if you have had previous hours and hours of machine learning theory but could never got a change or figure out how to implement and solve data science problems with it.

    The approach in this course is very practical and we will start everything from very scratch. We will immediately start coding after a couple of introductory tutorials and we try to keep the theory to bare minimal. All the coding will be done in MATLAB which is one of the fundamental programming languages for engineer and science students and is frequently used by top data science research groups world wide.

    Below is the brief outline of this course.

    1: Introduction to course

    In this section we spend some time talking about the topics you’ll learn, the approach of learning used in the course, essential details about MATLAB to get you started. This will give you an idea of what to expect from the course.

    2: Data preprocessing (Brief videos)

    We need to prepare and preprocess our data before applying Data Science algorithms and techniques. This section discusses the essential preprocessing techniques and discuses the topics such as getting rid of outliers, dealing with missing values, converting categorical data to numerical form, and feature scalling.

    3: Classification Algorithms in MATLAB

    Classification algorithms is an important class of Data Science algorithms and is a must learn for every data scientist. This section provides not only the intuition behind some of the most commonly used classification algorithm but also provides there implementation in MATLAB. The algorithms that we cover are

    a). K-Nearest Neighbor

    b). Naïve Bayesain

    c). Support Vector Machine

    d). Decision Trees

    e). Discriminant Analysis

    f). Ensembles

    In addition to these we also cover how to evaluate the performance of classifiers using different metrics.

    4: Clustering Algorithms in MATLAB

    This section introduces some of the commonly used clustering algorithms alongside with their intuition and implementation in MATLAB. We also cover the limitations of clustering algorithms by looking at their performance when the clusters are of different sizes, shapes and densities. The algorithms we cover in this section are

    a). K-Means

    b). Mean Shift

    c). DBSCAN

    d). Hierarchical Clustering

    In the same section, we also cover practical application of the clustering algorithms by looking at the applications of image compression and sentence grouping. This section provides some intuition regarding the strengths of clustering in real life data analysis tasks.

    5: Dimensionality Reduction

    Dimensionality reduction is an important branch of algorithms in Data Science. In this section we show how to reduce the dimensions for a specific Data Science problems so that the visualization becomes easy. We cover the PCA algorithm in this section.

    6: Project: Malware Analysis

    In this section we provide a detailed project on malware analysis from one of our recent research paper. We provide introductory videos on how to complete the project. This will provide you with some hands on experience for analyzing Data Science problems.

    7: Data preprocessing (Detailed Videos)

    In this section we dive deep into the topic of data preprocessing and cover many interesting topics. The topic in this section include

    a). Dealing with missing data using

    b). Deleting strategies

    c). Using mean and mode

    d). Radom values for handling missing data

    e). Class based strategies

    f). Considering as a special value

    Dealing with Categorical Variables using the

    a). One hot encoding

    b). Frequency based encoding

    c). Target based encoding

    d). Encoding in the presence of an order

    Outlier Detection using

    a). 3 sigma rule with

    b). Box plot rule

    c). Histogram based rule

    d). Local outlier factor

    e). Outliers in categorical variable

    f). Feature Scaling and Data Discretization

     

    For more inputs on Machine Learning for Data Science using MATLAB you can connect here.
    Contact the L&D Specialist at Locus IT.

    Reviews

    There are no reviews yet.

    Be the first to review “Machine Learning for Data Science using MATLAB”

    Your email address will not be published.