Page 271 - AI Ver 3.0 class 10_Flipbook
P. 271
Transform Widgets
Transform Widgets are used for modifying and transforming data in your machine learning workflow. They allows
users to apply various preprocessing and feature engineering techniques to manipulate datasets, preparing them
for advanced analysis or model building. In short, we can say that these widgets help perform different operations
on data.
Some of these Transform widgets are:
• Data Sampler Widget: It is used to sample or split a dataset into smaller subsets. It is useful when you
want to work with a random subset of your data for training or testing purposes, or when you want to perform
cross-validation.
• Select Columns Widget: It is used to select specific attributes (columns) from a
dataset that you want to work with, while excluding others. This is helpful for feature
selection, when you want to include only the relevant features for training a model.
• Impute Widget: It is used to handle missing data in your dataset. Missing values are
a common problem in real-world datasets, and imputation is the process of replacing
these missing values with some form of estimate (e.g., mean, median, mode).
• Discretize Widget: It is used to convert continuous data (like numbers) into categories
(like groups or ranges), making it easier for certain machine learning models to process
and understand the data.
Visualize Widgets
Visualize Widgets create charts and graphs that allow you to visualise patterns, distributions, and relationships
in your data. Visualization is an essential part of the data exploration process, as it can reveal insights that are not
easily apparent from raw data alone.
Some of these Visualize widgets are:
• Tree Viewer Widget: It is used to visualize decision trees. A decision tree is a popular
type of machine learning model used for making predictions. It’s like a flowchart that
helps make decisions based on different rules or conditions.
• Box Plot Widget: It is used to show the distribution of a continuous numerical feature.
It displays the median, quartiles, and outliers, giving a clear picture of how the data is
spread.
• Scatter Plot Widget: It is used to visualize the relationship between two continuous
numerical features. This widget is ideal for detecting correlations, trends, and clustering
within the data.
• Bar Plot Widget: It is used to visualize the distribution of categorical data by displaying
the frequency or count of categories. It provides a clear and simple way to compare the
size of different categories.
Model Widgets
Model widgets are used to create, train, and evaluate machine learning models. These widgets are essential for
building and testing models that make predictions based on the input data. For example, you might use a model
to predict whether someone will like a product, or how much a house might cost based on its features.
Statisical Data (Practical) 269

