
Explore a machine learning workflow to classify stars, galaxies, and quasars, from dataset understanding and preprocessing to model building with decision tree and logistic regression, and evaluate accuracy.
Discover the definitions and morphologies of stars, galaxies, and quasars, and learn to predict whether a night-sky dot is a star, galaxy, or quasar from the dataset.
Explore the Sloan Digital Sky Survey Data Release 14 dataset, a catalog of 10,000 observations with 17 features and a star/galaxy/quasar label, featuring photometric and spectral data.
Import libraries for data handling, visualization, and model building using os, numpy, pandas, matplotlib, seaborn, TensorFlow, and sklearn, and review the SDS Dr14 dataset for classifying stars, galaxies, and quasars.
Explore the dataset with pandas, drop unnecessary columns, and use head, shape, describe, and info to verify a complete 10,000×16 data table before proceeding to label encoding.
Encode the class column using sklearn's LabelEncoder to convert star, galaxy, and quasar labels into numeric values, enabling downstream model training with fit and transform.
Split the dataset into training and testing sets (70/30), then scale with a standard scaler. Avoid data leakage by fitting only on X_train and transforming X_test.
Use seaborn to visualize class distributions (galaxies, quasars, and stars) with count plots and explore relationships with pair plots across ultraviolet, green, red, and infrared bands.
Explore how to build a decision tree classifier with ID3, understand entropy and information gain, and implement with sklearn to classify stars, galaxies, and quasars.
Explore logistic regression for multi-class classification, predicting star, galaxy, or quasar with Python and sklearn, using the sigmoid function to convert outputs to class probabilities and threshold values.
Get a quick overview of the k-nearest neighbors classifier, build it with sklearn using three neighbors, and explore its strengths while noting the exoplanets video for deeper ideas.
Train and evaluate three ML classifiers by fitting to the training data and predicting on the test set, then compare predictions to the original dataset to assess accuracy amid imbalance.
evaluate ML models with the classification report, focusing on precision, recall, and f1 score, and address dataset imbalance among stars, galaxies, and quasars.
Concludes the stars, galaxies and quasars prediction project by outlining data leakage fixes, eda with pandas, cleaning, preprocessing, and handling imbalanced data with f1 score.
Welcome to "Project - Predicting Stars, Galaxies, and Quasars with ML", a specialized course where the realms of astronomy and machine learning intersect to create a unique and compelling learning experience.
About This Course: Are you fascinated by the universe and the treasures it holds? Do you have a passion for data science and its limitless possibilities? This course is designed to bring these two worlds together by teaching you how to apply machine learning techniques to classify cosmic bodies such as stars, galaxies, and quasars.
What You'll Learn:
Foundational Astronomy: Get to grips with the astronomical basics necessary for understanding the data you’ll be working with, including the characteristics of different celestial objects.
Data Handling Skills: Learn how to preprocess, clean, and manipulate large datasets typical of astronomical data collection efforts.
Machine Learning Techniques: Dive deep into machine learning algorithms, understanding how they can be trained to categorize vast and complex cosmic datasets.
Model Evaluation: Master the techniques for validating the performance of your machine learning models, ensuring accurate and reliable classification results.
Real-world Application: Apply your skills to a project that mirrors real-world tasks faced by astrophysicists and data scientists in the field.
Who This Course Is For: This course is ideal for data scientists, machine learning enthusiasts, and anyone with a curiosity about astronomy and how modern technology can be used to decipher it. Whether you’re looking to apply machine learning to a new domain or you’re an amateur astronomer eager to explore how data science can further your observations, this course will provide the knowledge and hands-on experience you’re looking for.
By the End of This Course: You will have completed a project that not only demonstrates your ability to handle complex datasets but also your capability to extract meaningful insights from the noise of the cosmos. You will have the confidence to tackle similar challenges, the skills to further your study, and the foundational knowledge to pursue more advanced work in either field.
Prepare to embark on a journey of discovery that will take you from the classroom to the cosmos, as you learn to harness the power of machine learning to unlock secrets of the universe that have puzzled humanity for centuries.
Enroll now and start your adventure into the cosmos with the power of machine learning!