Applying the MLP classifier class on the fertility set

The MLP classifier in scikit-learn has around 20 different hyperparameters associated with the architecture or the learning process, which can be altered in order to modify the training process of the network. Fortunately, all of these hyperparameters have set default values, which allows us to run an initial model without much effort. The results from this model can then be used to tune the hyperparameters as required.

Importing libraries and packages

# Mathematical operations and data manipulation
import pandas as pd

# Model
from sklearn.neural_network import MLPClassifier
from sklearn.metrics import accuracy_score

# Warnings
import warnings

warnings.filterwarnings("ignore")

Set paths

# Path to datasets directory
data_path = "./datasets"
# Path to assets directory (for saving results to)
assets_path = "./assets"

Loading dataset

The Fertility dataset aims to determine whether the fertility level of an individual has been affected by their demographics, their environmental conditions, and their previous medical conditions

dataset = pd.read_csv(f"{data_path}/fertility_Diagnosis.csv", header=None)

Exploring dataset

# Shape of the dataset
print("Shape of the dataset: ", dataset.shape)
# Head
dataset

Shape of the dataset:  (100, 10)

	0	1	2	3	4	5	6	7	8	9
0	-0.33	0.69	0	1	1	0	0.8	0	0.88	N
1	-0.33	0.94	1	0	1	0	0.8	1	0.31	O
2	-0.33	0.50	1	0	0	0	1.0	-1	0.50	N
3	-0.33	0.75	0	1	1	0	1.0	-1	0.38	N
4	-0.33	0.67	1	1	0	0	0.8	-1	0.50	O
...	...	...	...	...	...	...	...	...	...	...
95	-1.00	0.67	1	0	0	0	1.0	-1	0.50	N
96	-1.00	0.61	1	0	0	0	0.8	0	0.50	N
97	-1.00	0.67	1	1	1	0	1.0	-1	0.31	N
98	-1.00	0.64	1	0	1	0	1.0	0	0.19	N
99	-1.00	0.69	0	1	1	0	0.6	-1	0.19	N

100 rows × 10 columns

Modelling

X = dataset.iloc[:, :9]
Y = dataset.iloc[:, 9]

model = MLPClassifier(max_iter=300, random_state=101)
model.fit(X, Y)

MLPClassifier(max_iter=300, random_state=101)

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

model = MLPClassifier(random_state=101, max_iter=1200)
model.fit(X, Y)

MLPClassifier(max_iter=1200, random_state=101)

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

# Testing by performing a prediction for a new instance with feature values
# −0.33 , 0.69 , 0 , 1 , 1 , 0 , 0.8 , 0 , 0.88 (first row X)
pred = model.predict([[-0.33, 0.69, 0, 1, 1, 0, 0.8, 0, 0.88]])
print(pred)

['N']

pred = model.predict(X)
score = accuracy_score(Y, pred)
print(score)

0.98