Error in Perceptron `fit`

[FrankC01]

OS: Big Sur Python (venv): 3.9.7 numpy: 1.21.2 pandas: 1.3.3

When executing fit in Chapter 2:

self.w_[0] += update
ValueError: setting an array element with a sequence

[rasbt]

Hi Frank, could you share the whole code you are using for chapter 2? Maybe there is a typo somehwere regarding calculating "update". For reference the relevant line is this: https://github.com/rasbt/python-machine-learning-book-3rd-edition/blob/master/ch02/ch02.py#L137

[FrankC01]

Hand typed from the book:

Perceptron

import numpy as np

class Perceptron(object):
    def __init__(self, eta=0.01, n_iter=50, random_state=1) -> None:
        self.eta = eta
        self.n_iter = n_iter
        self.random_state = random_state

    def fit(self, X, y):
        rgen = np.random.RandomState(self.random_state)
        self.w_ = rgen.normal(loc=0.0, scale=0.01, size=1 + X.shape[1])
        self.errors_ = []

        for _ in range(self.n_iter):
            errors = 0
            for xi, target in zip(X, y):
                update = self.eta * (target - self.predict(xi))
                self.w_[1:] += update * xi
                self.w_[0] += update
                errors += int(update != 0.0)
            self.errors_.append(errors)
        return self

    def net_input(self, X):
        return np.dot(X, self.w_[1:]) + self.w_[0]

    def predict(self, X):
        return np.where(self.net_input(X) >= 0.0, 1, -1)

Driver

#!/usr/bin/env python3
# -*- coding: utf-8; py-indent-offset:4 -*-

import os
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from perceptron import Perceptron

def perceptron():
    s = os.path.join('https://archive.ics.uci.edu', 'ml',
                     'machine-learning-databases', 'iris', 'iris.data')
    print(f"Url : {s}")
    df = pd.read_csv(s, header=None, encoding='utf-8')
    # print(f"{df.tail()}")
    # y df is the first 100 using 4 columns
    y = df.iloc[0:100, 4].values
    # print(f"{y}")
    y = np.where(y == 'Iris-setosa', -1, 1)
    # print(f"{y}")
    # X df is the lengths of sepal and petal (Cols 0 and 2)
    X = df.iloc[0:100, [0, 2]]
    #
    plt.scatter(X.iloc[:50, 0], X.iloc[:50, 1],
                color='red', marker='o', label='setosa')
    plt.scatter(X.iloc[50:100, 0], X.iloc[50:100, 1],
                color='blue', marker='x', label='versicolor')
    plt.xlabel('sepal length [cm]')
    plt.ylabel('petal length [cm]')
    plt.legend(loc='upper left')
    # plt.show()

    # Do some ML!!!
    ppn = Perceptron(eta=0.1, n_iter=10)
    ppn.fit(X, y)
    plt.plot(range(1, len(ppn.errors_) + 1), ppn.errors_, marker='o')
    plt.xlabel('Epochs')
    plt.ylabel('Number of updates')
    plt.show()

if __name__ == '__main__':
    perceptron()

[rasbt]

That's because the X array is a pandas DataFrame that creates the mismatch. If you change

ppn.fit(X, y)

to

ppn.fit(X.values, y)

it should work.

[FrankC01]

That did it but the book and the example do not show that.

[rasbt]

In the book, it was done earlier when preparing the X numpy array:

You could also do that in your code above, but then you have to adjust the scatter function in your code, which uses .iloc, so I suggested in your code just to modify the part

ppn.fit(X.values, y)

[FrankC01]

Ahh, correct... I did not have X = df.ilog[0:100, [0,2]].values but X = df.ilog[0:100, [0,2]] and because of that I was getting errors on the scatter which fixed with iloc.

Clearly I'm n00b for pandas, numpy and NL :)... what better way to spend on a rainy day!

Thanks!

[rasbt]

No worries, these things are really easy to overlook :). Glad it was an easy fix though!