How to fix invalid index to scalar variable Using NLTK | Python

Question

I am trying to run a NLP Sentiment Analysis using Naives Bayes algorithm. But I encounter the following error in the final stage, In [13]:

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
Cell In[13], line 1
----> 1 acc = accuracy(X_train, Y_train, X_test, Y_test)
      2 acc

Cell In[12], line 5, in accuracy(X_train, Y_train, X_test, Y_test)
      2 pred = []
      4 for i in range(X_test.shape[0]):
----> 5     p = predict(X_train,Y_train, X_test[i])
      6     pred.append(p)
      8 Y_pred = np.array(pred)

Cell In[11], line 10, in predict(X_train, Y_train, X_test)
      8 likelihood = 1.0
      9 for fea in range(n_features):
---> 10     cond  = cond_probab(X_train, Y_train, fea, X_test[fea], label)
     11     likelihood = likelihood * cond
     13 prior = prior_probab(Y_train, label)

IndexError: invalid index to scalar variable.

The whole code in Python (in Jupyter Notebook) is the following:

In 1: Import

import pandas as pd
import numpy as np
from sklearn.preprocessing import LabelEncoder
from sklearn.model_selection import train_test_split

In [2]: Import dataset from Kaggle: Mushroom Classification

df = pd.read_csv("C:/.../mushrooms.csv")

In [3 - 5]: Convert values into numbers

le = LabelEncoder()
df_encoded = df.apply(le.fit_transform, axis = 0)
df =  df_encoded.values

In [6 - 7]: Define X and Y datasets

X = df[: , 1: ]
Y = df[:,0] 

In [8]: Split dataset

X_train, Y_train, X_test, Y_test = train_test_split(
    X, Y, test_size = 0.20, random_state = 42)

In [9]: Create a function to find Prior Probability:

def prior_probab(Y_train, label):
    m = Y_train.shape[0]
    s = np.sum(Y_train == label)
    
    return m/s

In [10]: Create Condition Probability to find Likelihood

def cond_probab(X_train, Y_train, feature_col, feature_val, label):
    
    X_filtered = X_train[Y_train == label]
    num = np.sum(X_filtered[:,feature_col] == feature_val)
    
    denom = X_filtered.shape[0]
    
    return float(num/denom)

In [11]: Create Prediction

def predict(X_train, Y_train, X_test):
    
    classes = np.unique(Y_train)
    n_features = X_train.shape[1]
    posterior_probab = []
    
    for label in classes:
        likelihood = 1.0
        for fea in range(n_features):
            cond  = cond_probab(X_train, Y_train, fea, X_test[fea], label)
            likelihood = likelihood * cond
            
        prior = prior_probab(Y_train, label)
        post = likelihood * prior
        
        posterior_probab.append(post)
        
        pred = np.argmax(posterior_probab)
        
        return pred

In [12]: Create Accuracy

def accuracy(X_train, Y_train, X_test, Y_test):
    pred = []
    
    for i in range(X_test.shape[0]):
        p = predict(X_train,Y_train, X_test[i])
        pred.append(p)
    
    Y_pred = np.array(pred)
    acc = np.sum(Y_pred == Y_test)/Y_pred.shape[0]
    
    return acc

In [13]: Result

acc = accuracy(X_train, Y_train, X_test, Y_test)
acc