Machine Learning

Making Predictions Based on Data Analysis

AI Programming & Machine Learning4.6(80)9h 54m52 lectures5 sections

What you'll learn

  • Define machine learning, its applications and its categories
  • Install Anaconda and get to know the Python language
  • Understand simple and multiple linear regression
  • Apply nonlinear regression
  • Evaluate a model and measure its performance
  • Classify with the nearest neighbors algorithm and the decision tree
  • Use the SVM algorithm and logistic regression
  • Cluster with the K-means algorithm and hierarchical clustering
  • Build recommendation systems

About this course

This course teaches Machine Learning using the Python language, meaning how to make the computer learn from data and make predictions. You start with a definition of machine learning, its applications and its categories, why we use Python, and installing Anaconda. Next you move on to supervised learning and get to know regression, simple and multiple linear regression and nonlinear regression, along with ways to evaluate a model and measure its performance. Then you learn classification through the nearest neighbors algorithm, the decision tree, logistic regression and the SVM algorithm, and clustering through the K-means algorithm and hierarchical clustering, and you finish with content-based and collaborative recommendation systems. Every topic comes with a theory explanation followed by hands-on coding on ready-made data files. This course suits beginners in data analysis and anyone who wants to enter the field of artificial intelligence, and it helps to have a basic idea of programming. By the end, you will be able to build machine learning models, evaluate them and use them to make predictions from your own data.

Expected outcomes

  • Ability to build machine learning models using modern tools and techniques.
  • Understanding how to apply supervised and unsupervised learning algorithms.
  • Ability to analyze data and use it to improve performance and predictions.
  • Creating customized recommendation systems for various practical applications.
  • Developing practical projects that make machine learning a part of your professional life.

Learner reviews

These reviews are from learners who bought the content and completed at least 50% of it — there are no reviews from anyone who hasn't gone through the material.
4.6
★★★★★
80 reviews
5★
50
4★
28
3★
2
2★
0
1★
0
  • ★★★★★عبدالرحمن ع.·منذ 4 شهرشاهد 100٪

    كورس فوق الممتاز وملم بكل جوانب تعلم الالة وكافية جدا لاي حد حابب يفهم ويدخل المجال

تم عرض كل التقييمات (1)

Course content

1

Introduction to Machine Learning

4 lectures
  1. Definition of Machine Learning and Its ApplicationsDefines Machine Learning as a branch of computer science where computers learn from data without explicit programming, and shows applications such as YouTube, Amazon and Netflix recommendations and spam filtering.9:33
  2. Machine Learning ClassificationsExplains the three categories of machine learning: supervised learning with regression and classification, unsupervised learning with clustering, dimension reduction and anomaly detection, and reinforcement learning such as Q-Learning.12:43
  3. Why Python Language?Explains why Python is used for machine learning and introduces the NumPy, SciPy, Matplotlib and Pandas libraries, covering what each offers for calculations, plotting and data handling.17:06
  4. Installing AnacondaShows how to download Anaconda from its website and install it, then open Anaconda Navigator and confirm that the installation succeeded and the tools appear.2:09
2

Supervised Learning - Regression

15 lectures
  1. Introduction to Supervised LearningAn introduction to Supervised Learning, which learns from examples of inputs and outputs, and its two forms, Classification and Regression, with examples and a preview of the next section.3:41
  2. What is Regression?Defines Regression and its prediction of a continuous value, distinguishes dependent from independent variables, and presents its types, Simple and Multiple, Linear and Non-Linear, with a car CO2 emission example.5:13
  3. Simple Linear RegressionExplains Simple Linear Regression with one independent variable, finding the line's coefficients theta 0 and theta 1 by minimizing the error between actual and predicted values, worked by hand to predict CO2 emission.25:43
  4. Practical ExampleA hands-on linear regression example using Anaconda and Spyder, where scikit-learn computes the equation coefficients theta 0 and theta 1 and predicts CO2 emissions from engine size.16:17
  5. 1FuelConsumptionCo2
  6. Model EvaluationExplains the two ways to evaluate a regression model, testing on the training data versus a train/test split, including how error is computed and the problem of overfitting.7:06
  7. Model Performance MeasurementExplains the metrics used to measure regression model performance: mean absolute error, mean squared error, root mean squared error, and relative absolute and relative squared error, with the formula of each.10:12
  8. Practical Application of Simple Linear RegressionA hands-on simple linear regression example with scikit-learn on CO2 emissions data, splitting it into train and test sets and evaluating the model with the R2 score.33:24
  9. Multiple Linear Regression (Part One)Introduces multiple linear regression, which differs from simple regression by using several independent variables to predict one dependent variable, with examples like CO2 emissions, student performance and blood pressure.6:19
  10. Multiple Linear Regression (Part Two)Explains the multiple linear regression equation and its theta coefficients, how error between predicted and actual values is computed, and the goal of finding coefficients that minimize that error.11:31
  11. Practical Application of Multiple Linear RegressionA hands-on multiple linear regression example with scikit-learn on CO2 emissions data, splitting the data into train and test sets, training the model and then predicting values.27:31
  12. 1china_gdp
  13. Questions about Multiple Linear RegressionAnswers common questions about multiple linear regression: when to use it instead of simple regression, how many independent variables to choose, how to handle text values, and linear versus non-linear relationships.9:03
  14. Non-Linear Regression (Part One)Distinguishes linear from non-linear regression equations and plots Python examples of a linear function, a cubic polynomial, a quadratic and a logarithmic function with added random noise.18:01
  15. Non-Linear Regression (Part Two)Applies non-linear regression to China's GDP over the years, using a sigmoid function to model the curve and tuning its parameters so it fits the data.29:22
3

Supervised Learning - Classification

17 lectures
  1. Introduction to ClassificationAn introduction to Classification, a type of supervised learning that builds a model assigning inputs to classes, such as labeling email as Spam or Not Spam, dog images, and medical diagnosis.6:16
  2. KNN Nearest Neighbors AlgorithmExplains the KNN nearest neighbors classification algorithm, which labels a new customer by the majority class among its K closest known cases, using a telecom services example, and raises the questions of choosing K and measuring similarity.5:47
  3. Identifying and Calculating ValuesAnswers two KNN questions: how to compute the distance between two cases using the distance formula over their features, and how to choose a suitable K by avoiding overfitting and underfitting through accuracy tests on held-out data.9:19
  4. Model Evaluation MetricsExplains model evaluation metrics computed on a test set, including accuracy, F1 score, the confusion matrix, precision and recall, and Log Loss for models that output probabilities.18:13
  5. Practical Application of the Nearest Neighbors AlgorithmA practical K-Nearest Neighbors example for classifying customers, including the train/test split, choosing the best K by comparing accuracy, and predicting the class of a new customer.32:45
  6. 1teleCust1000t
  7. Decision Tree AlgorithmIntroduces the Decision Tree classification algorithm, explaining nodes, branches and leaves through a patient and drug example, and the steps of building it by choosing key features and splitting data.7:28
  8. Building a Decision TreeExplains how a decision tree is built by choosing the attribute that reduces impurity, defining entropy and Information Gain to decide the best split at each node.18:37
  9. Practical Application of the Decision TreeA practical decision tree example that classifies drug types from patient data, including converting text values to numbers, training and testing the model, and measuring its accuracy.25:26
  10. 1drug200
  11. Introduction to Logistic RegressionIntroduces Logistic Regression as a classification algorithm that predicts a class and its probability rather than a number, with example problems, when to use it, and how it compares to linear regression.9:53
  12. Understanding the Model Training PhaseExplains training a logistic regression model, from the cost function and finding the best theta values with gradient descent to the steps of training, evaluation and prediction.11:17
  13. Practical Application of the Decision Tree and Model TrainingA practical logistic regression example on telecom customer data to predict service cancellation, covering feature selection, data splitting, computing probabilities and measuring accuracy.20:54
  14. 1ChurnData
  15. SVM AlgorithmExplains the Support Vector Machine classification algorithm: transforming data with kernel functions so a plane can separate it, finding the best separator via support vectors, and applications like images, spam and sentiment analysis.13:50
  16. SVM Practical ApplicationA practical Support Vector Machine example that classifies benign and malignant tumors from cell features, including plotting the data, training the model with a kernel function and testing it.28:01
  17. 1cell_samples
4

Unsupervised Learning - Clustering

10 lectures
  1. Introduction to ClusteringAn introduction to Clustering, the first type of unsupervised learning, and how it splits data into groups with similar traits, such as customer segmentation, raising the question of how it differs from classification.7:07
  2. Difference Between Clustering and ClassificationClarifies the difference between Clustering and Classification: the first is unsupervised on unlabeled data, the second supervised on labeled data, and presents the three types of clustering.10:41
  3. KM AlgorithmExplains the K-Means clustering algorithm for splitting customers into K groups: picking random centers, computing distances, assigning each point to its nearest center, then updating the centers, and warns about the effect of random initialization.12:47
  4. Measuring Algorithm AccuracyExplains the two ways to measure K-Means accuracy, external and internal approaches, and how to choose the number of clusters K using the Elbow Point.3:06
  5. Practical Application of the AlgorithmA practical K-Means example that clusters customer data, covering data preparation and scaling, model training, assigning a cluster to each customer, and plotting the clusters in 3D.19:58
  6. 1cust
  7. Introduction to Hierarchical ClusteringIntroduces hierarchical clustering and its divisive and agglomerative types, using a cities example to show how clusters merge by smallest distance and are drawn as a dendrogram.9:09
  8. Hierarchical and K-Means ClusteringExplains the bottom-up agglomerative clustering steps and the distance matrix, plus the ways to measure distance between clusters: single, complete, average and centroid linkage.4:43
  9. Agglomerative Practical Application of the AlgorithmA practical Agglomerative clustering example that groups cars by similar features, covering data cleaning, computing the distance matrix and displaying the clusters as a dendrogram.22:55
  10. 1cars_clus
5

Recommender System

6 lectures
  1. Introduction to Recommendation SystemsAn introduction to Recommendation Systems and their value on platforms like Amazon and Netflix, covering the two types, Content-Based and Collaborative Filtering, and the two implementation approaches, Memory-Based and Model-Based.5:58
  2. Content-Based SystemExplains the Content-Based recommender system and how it builds a User Profile from a person's ratings and the genres of films watched to suggest similar films, notes its weakness and leads into collaborative filtering.11:46
  3. Practical Application of ContentA hands-on Python build of a content-based movie recommender from movies, ratings and tags files, converting genres into a matrix and building a user profile to rank film suggestions.21:40
  4. Project Files
  5. Collaborative-Based SystemExplains collaborative filtering in recommender systems, user-based and item-based, with a movie example that computes similarity between users to predict ratings for unseen films.11:45
  6. Practical Application on Recommendation Systems

Instructor

Eng. Amr Abdel Fattah

Eng. Amr Abdel Fattah

Computer systems engineer specializing in mobile app and website development, with experience in creating popular platforms and web applications.
49.1K views20 courses

This course is part of the diploma

Related courses