Suggested Table Structure for ML Models Project
Your final results should be presented in a table format that allows for clear comparison of models
and datasets.
Dataset Name ML Model Hyperparameters Accuracy F1-Score MSE
load_iris() Logistic Regression C=1.0 0.96 0.96
load_iris() Decision Tree max_depth=5 0.98 0.98
load_boston() Linear Regression - - 21.8
make_blobs() K-Means n_clusters=3 0.85 -
List of Sklearn Models
Name Sklearn Function (Class Name) Type
Linear Regression LinearRegression Regression
Logistic Regression LogisticRegression Classification
Support Vector Machine (SVC) SVC Classification
Support Vector Machine (SVR) SVR Regression
Decision Tree DecisionTreeClassifier Classification
Decision Tree Regressor DecisionTreeRegressor Regression
Random Forest RandomForestClassifier Classification
Random Forest Regressor RandomForestRegressor Regression
K-Nearest Neighbors (KNN) KNeighborsClassifier Classification
KNN Regressor KNeighborsRegressor Regression
Gaussian Naive Bayes GaussianNB Classification
Multinomial Naive Bayes MultinomialNB Classification
K-Means KMeans Clustering
Agglomerative Clustering AgglomerativeClustering Clustering
AdaBoost AdaBoostClassifier Classification
Bagging BaggingRegressor Regression
Toy Datasets
Name SklearnFunction Type Description
Classic dataset for simple classification,
Iris load_iris() Classification containing measurements of three species
of Iris flowers (150 samples, 4 features).
Small image dataset of handwritten digits
Digits load_digits() Classification from 0 to 9, for pattern recognition (1797
samples, 64 features).
Results of a chemical analysis of wines
grown in the same region in Italy but derived
Wine load_wine() Classification
from three different cultivars (178 samples,
13 features).
Wisconsin Diagnostic Breast Cancer
(WDBC) dataset for binary classification
Breast Cancer load_breast_cancer() Classification
(malignant or benign) (569 samples, 30
features).
Ten baseline variables, age, sex, body mass
index, average blood pressure, and six
blood serum measurements, for 442
Diabetes load_diabetes() Regression
diabetes patients, used to predict a
quantitative measure of disease
progression.
Physical exercise and physiological data for
Multi-target
Linnerud load_linnerud() 20 subjects, often used for demonstrating
Regression
multi-target prediction.
Generates isotropic Gaussian clusters for
Generated
clustering or classification tasks, where the
Blobs make_blobs() (Clustering/
number of samples, centers, and features
Classification)
can be customized.
Generated Generates a two-interleaving half-moon
Moons make_moons() (Classification/ dataset, useful for visualizing non-linear
Clustering) separation boundaries.
Generated Generates two concentric circles, also used
Circles make_circles() (Classification/ for visualizing non-linear separation
Clustering) problems.
Generates a random multi-class
Generated classification problem with customizable
Classification make_classification()
(Classification) parameters like the number of features,
informative features, and classes.
Name Use Type Common Source Description
Titanic Kaggle / GitHub Predict survival on the Titanic based on
Classification
Survival Repositories features like age, class, sex, and fare.
TensorFlow
Large dataset of $70,000$ handwritten digits
Image Datasets /
MNIST ($0$ through $9$) used for training image
Classification PyTorch / GitHub
processing systems.
Repositories
Scikit-learn Predict the median house value for districts in
California
Regression fetchers / UCI ML California using features from the $1990$
Housing
Repository Census.
TensorFlow
IMDB Movie Text Dataset for sentiment analysis, containing
Datasets / GitHub
Reviews Classification $50,000$classified reviews (positive/negative).
Repositories
UCI ML
Adult (Census Predict whether a person's income exceeds
Classification Repository / GitHub
Income) $50K/year based on census data.
Repositories
Credit Card
Kaggle / GitHub Highly imbalanced dataset of credit card
Fraud Classification
Repositories transactions to detect fraudulent ones.
Detection
Remember to properly preprocess data (e.g., scaling features for SVM/KNN) and split your data
(train/test) for supervised learning tasks.