Scribd Logo
Upload

Welcome to Scribd!

  • 27 views

    Machine Learning

    Uploaded by

    Asmaa Nouali
    Machine learning is a category of algorithms that allows systems to learn from data patterns without being explicitly programmed. There are two main types of machine learning: supervised learning, which uses labeled training data to predict outputs, and unsupervised learning, which finds hidden patterns in unlabeled data. Common supervised learning algorithms include regression, which predicts continuous values, and classification, which predicts categorical labels. Regression techniques include linear and multi-linear regression, while classification includes logistic regression and K-nearest neighbors (KNN) algorithms. KNN is a simple algorithm that classifies points based on their closest neighbors.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Machine Learning

    Uploaded by

    Asmaa Nouali
    27 views37 pages
    Machine learning is a category of algorithms that allows systems to learn from data patterns without being explicitly programmed. There are two main types of machine learning: supervised learning, which uses labeled training data to predict outputs, and unsupervised learning, which finds hidden patterns in unlabeled data. Common supervised learning algorithms include regression, which predicts continuous values, and classification, which predicts categorical labels. Regression techniques include linear and multi-linear regression, while classification includes logistic regression and K-nearest neighbors (KNN) algorithms. KNN is a simple algorithm that classifies points based on their closest neighbors.

    Original Title

    Machine Learning (3)

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Machine learning is a category of algorithms that allows systems to learn from data patterns without being explicitly programmed. There are two main types of machine learning: supervised learning, which uses labeled training data to predict outputs, and unsupervised learning, which finds hidden patterns in unlabeled data. Common supervised learning algorithms include regression, which predicts continuous values, and classification, which predicts categorical labels. Regression techniques include linear and multi-linear regression, while classification includes logistic regression and K-nearest neighbors (KNN) algorithms. KNN is a simple algorithm that classifies points based on their closest neighbors.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    27 views37 pages

    Machine Learning

    Uploaded by

    Asmaa Nouali
    Machine learning is a category of algorithms that allows systems to learn from data patterns without being explicitly programmed. There are two main types of machine learning: supervised learning, which uses labeled training data to predict outputs, and unsupervised learning, which finds hidden patterns in unlabeled data. Common supervised learning algorithms include regression, which predicts continuous values, and classification, which predicts categorical labels. Regression techniques include linear and multi-linear regression, while classification includes logistic regression and K-nearest neighbors (KNN) algorithms. KNN is a simple algorithm that classifies points based on their closest neighbors.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 37

    What is

    Machine
    learning
    And why is everyone talking
    about it?
    Machine learning

    Machine learning is a category of algorithm that allows systems to


    explore data in order to perform a specific task effectively without
    being explicitly programmed. Instead, it relies on patterns and
    relationships.
    How is this possible? How does this mysterious process
    happen? That’s what we’ll be seeing

    in this speech but first let’s take a look


    at machine learning applications
    and its different types.
    Types of ML
    Supervised learning is where you:
    -Have labeled data which means we already know the corresponding output for
    each input.
    -Use an algorithm to learn the hidden patterns between the inputs and the output.
    -Tend to model these hidden patterns by a mapping function.
    We’ll define it by saying that supervised learning is based on the concept of
    teaching a machine learning algorithm how to predict using already divided
    (labeled) data.
    REGRESSION CLASSIFICATION
    In Regression:
    The output variable is a continuous value.
    The input variable is a continuous value.
    The mapping function is a mathematical function.
    EXAMPLE
    To predict the salary of an employee, let’s use a regression algorithm

    Linear Regression???

    Prediction Errors

    error = y - y_predicted
    Loss/Cost functions

    Multi-Linear Regression

    Multi-linear regression
    Model

    Two possible solutions :

    Overfitting Underfitting

    In regression problems, we can


    use linear, multilinear and
    polynomial regression.
    Graphical plots help us choose the type of regression.
    Regressions are sensitive to outliers.
    In classification :
    The output variable is a categorical variable.

    The mapping function is based on probabilities or the


    measure of similarities between a pair of objects.
    Logistic Regression

    EXAMPLE
    We can use it To predict if a candidate is getting admitted or not (yes or no)

    Sigmoid function???

    Confusion matrix???

    Logistic regression is used for binary classification problems with two possible outcomes.
    Logistic regression uses a sigmoid function as a linear equation to predict a value.
    To measure model performance, we use a confusion matrix where output can be two or more
    classes.
    KNN

    ~As known as the lazy learner~


    KNN: Non-Parametric
    KNN steps
    For a given data point x of an unknown class:
    Introduce the value of K.
    Calculate the distance between x and all the data points in
    the training data.
    Model picks K entries in the database which are closest to
    the new data point.
    Then it takes the majority vote i.e the most common
    class/label among those K entries and that will be the class of
    the new data point.
    KNN steps
    KNN Pros KNN Cons
    No assumptions about data. For
    01 example: unlike the ones imposed by the 01 K-NN needs homogeneous feature as
    features must have the same scale.
    linear model.
    K-NN might be very easy to implement but as the
    02 Simple algorithm: easy to
    understand and interpret. 02 dataset grows, the efficiency and/or the speed of
    the algorithm decline very fast.
    Outlier sensitivity: K-NN algorithm is very
    03 No training needed.
    03 sensitive to outliers as it simply chooses the
    neighbors based on distance criteria.
    Imbalanced data causes problems: For example, if
    04 Variety of distance functions. 04 we have two classes and one class is dominant
    more than the other, it will eventually represent
    80% of the dataset.

    KNN is a non-parametric machine learning algorithm.


    The purpose of KNN is to predict the classification of a new
    random point compared to other features(neighbors).
    To find the best k value, we apply the elbow method.
    Thank You

    You might also like