Example: Predicting Iris Flower Class
Data
- File: iris_data.csv
- Citation: UCI Iris
- data.txt
Network
- Four input features
- Two hidden layers
- Three output units: one for each class, encoding the probability for each class
- network.txt
Training Details
Details
This model predicts the type of iris based on four observed features. The desired output is a single sparse categorical value (translated from a string to a natural number)
Experiment Suggestion
Try changing the hidden layers and numbers of units to see if the accuracy changes
Scikit-Learn Models
- For these models, use experiment_skl.txt
- network_skl1.txt: SGDClassifier with Polynomial Feature preprocessing
- network_skl2.txt: Decision tree classifier
- network_skl3.txt: Random forest classifier