Scribd Logo
Upload

Welcome to Scribd!

  • 23 views

    Pca 1692550768

    Uploaded by

    kan luc N'guessan
    PCA is an unsupervised machine learning algorithm used for dimensionality reduction. It works by identifying principal components as linear combinations of the original variables that explain the maximum variance in the data. The algorithm standardizes data, computes the covariance matrix, calculates eigenvectors and eigenvalues to identify principal components, and projects the data onto these components to reduce dimensionality while retaining variation. PCA is useful for data visualization, compression, and preprocessing for machine learning models.

    Copyright:

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

    Pca 1692550768

    Uploaded by

    kan luc N'guessan
    23 views13 pages
    PCA is an unsupervised machine learning algorithm used for dimensionality reduction. It works by identifying principal components as linear combinations of the original variables that explain the maximum variance in the data. The algorithm standardizes data, computes the covariance matrix, calculates eigenvectors and eigenvalues to identify principal components, and projects the data onto these components to reduce dimensionality while retaining variation. PCA is useful for data visualization, compression, and preprocessing for machine learning models.

    Original Title

    PCA_1692550768

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    PCA is an unsupervised machine learning algorithm used for dimensionality reduction. It works by identifying principal components as linear combinations of the original variables that explain the maximum variance in the data. The algorithm standardizes data, computes the covariance matrix, calculates eigenvectors and eigenvalues to identify principal components, and projects the data onto these components to reduce dimensionality while retaining variation. PCA is useful for data visualization, compression, and preprocessing for machine learning models.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    23 views13 pages

    Pca 1692550768

    Uploaded by

    kan luc N'guessan
    PCA is an unsupervised machine learning algorithm used for dimensionality reduction. It works by identifying principal components as linear combinations of the original variables that explain the maximum variance in the data. The algorithm standardizes data, computes the covariance matrix, calculates eigenvectors and eigenvalues to identify principal components, and projects the data onto these components to reduce dimensionality while retaining variation. PCA is useful for data visualization, compression, and preprocessing for machine learning models.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 13

    Principal

    Component
    Analysis
    What is Principal
    Component Analysis?

    Principal Component Analysis (PCA) is an


    unsupervised learning algorithm that is used for
    dimensionality reduction in machine learning.

    PCA is used to reduce the dimensionality of a


    large dataset while retaining as much of the
    original variation as possible.
    How PCA works?

    PCA works by identifying a set of new variables,


    called principal components, that are linear
    combinations of the original variables, and then
    projecting the data onto these new variables.

    The first principal component is chosen to explain


    the maximum amount of variation in the data,
    and each subsequent component is chosen to
    explain the remaining variation in order of
    decreasing importance. The principal
    components are chosen such that they are
    orthogonal (i.e., uncorrelated) to each other.
    PCA Algorithm Steps

    1. Standardize the data: PCA assumes that the


    data is standardized (i.e., centered at zero and
    scaled to have unit variance). This step involves
    subtracting the mean of each variable from each
    observation and dividing it by the standard
    deviation of each variable.

    2. Compute the covariance matrix: The


    covariance matrix measures the pairwise
    covariances between all pairs of variables in the
    data. This matrix is used to compute the principal
    components.
    PCA Algorithm Steps

    3. Compute the eigenvectors and eigenvalues


    of the covariance matrix: The eigenvectors of
    the covariance matrix represent the directions in
    which the data varies the most, while the
    eigenvalues represent the amount of variance
    explained by each eigenvector.
    PCA Algorithm Steps

    4. Select the principal components: The


    eigenvectors are sorted in descending order of
    their corresponding eigenvalues, and the top k
    eigenvectors are selected as the principal
    components. The number of principal
    components to select depends on the amount of
    variance that needs to be explained and the
    desired level of dimensionality reduction.
    PCA Algorithm Steps

    5. Project the data onto the principal


    components: The original data is then projected
    onto the new set of k principal components. This
    involves computing the dot product of the original
    data matrix with the matrix of selected
    eigenvectors.

    6. Analyze the results: The resulting matrix of


    projected data can be analyzed using standard
    statistical techniques. The principal components
    themselves can also be examined to gain insight
    into the underlying structure of the data.
    Why it is useful?

    PCA can be used for a variety of purposes,


    including data visualization, data compression,
    and data pre-processing for machine learning
    algorithms.
    By reducing the dimensionality of the data, PCA
    can make it easier to analyze and interpret large
    datasets, and can also help to remove noise and
    redundancy in the data.
    Advantages

    1. Easy to calculate and compute.


    2. Speeds up machine learning computing
    processes and algorithms.
    3. Prevents predictive algorithms from data
    overfitting issues.
    4. Increases performance of ML algorithms by
    eliminating unnecessary correlated variables.
    5. Principal Component Analysis results in high
    variance and increases visualization.
    6. Helps reduce noise that cannot be ignored
    automatically.
    Disadvantages

    1. Sometimes, PCA is difficult to interpret. In rare


    cases, you may feel difficult to identify the most
    important features even after computing the
    principal components.

    2. You may face some difficulties in calculating


    the covariances and covariance matrices.

    3. Sometimes, the computed principal


    components can be more difficult to read rather
    than the original set of components.
    Follow #DataRanch on LinkedIn
    for more...
    Follow #DataRanch on LinkedIn
    for more...
    info@dataranch.org

    linkedin.com/company/dataranch

    You might also like