machine-learning-challenge

Machine Learning Homework - Exoplanet Exploration

Over a period of nine years in deep space, the NASA Kepler space telescope has been out on a planet-hunting mission to discover hidden planets outside of our solar system.

To help process this data, you will create machine learning models capable of classifying candidate exoplanets from the raw dataset.

In this homework assignment, we are to:

1. Preprocess the raw data
2. Tune the models
3. Compare two or more models

---

Instructions

Preprocess the Data

• Preprocess the dataset prior to fitting the model.
• Perform feature selection and remove unnecessary features.
• Use MinMaxScaler to scale the numerical data.
• Separate the data into training and testing data.

Tune Model Parameters

• Use GridSearch to tune model parameters.
• Tune and compare at least two different classifiers.

---

Resources

• Exoplanet Data Source

• Scikit-Learn Tutorial Part 1

• Scikit-Learn Tutorial Part 2

• Grid Search

---

Tools / Technologies

• Python
• SVM (Machine Learning Model)
• Random forest (Machine Learning Model)

Summary

Random Forest is better than SVM, even without GridSearchCV.

---

SVM

	Before CV	After CV
Training Score	0.85026	0.88777
Testing Score	0.83898	0.88106

---

	precision	recall	f1-score	support
CANDIDATE	0.85	0.65	0.73	523
CONFIRMED	0.75	0.87	0.81	594
FALSE POS	0.98	1.00	0.99	1069
micro avg	0.88	0.88	0.88	2186
macro avg	0.86	0.84	0.84	2186
weighted avg	0.88	0.88	0.88	2186

---

Random Forest

	Before CV	After CV
Training Score	0.99497	1.0
Testing Score	0.87740	0.89616

---

	precision	recall	f1-score	support
CANDIDATE	0.84	0.73	0.78	523
CONFIRMED	0.81	0.86	0.83	594
FALSE POS	0.97	1.00	0.98	1069
micro avg	0.90	0.90	0.90	2186
macro avg	0.87	0.86	0.87	2186
weighted avg	0.89	0.90	0.89	2186