This application uses machine learning techniques to classify emails as spam or not spam. It leverages the Scikit-learn library for building the classification model, pandas for data manipulation, and PyQt5 for the graphical user interface.
The basic UI showing Email form.
The next screenshots show emails which are classified as spam and not spam.
- Train a spam classifier using a dataset of labeled emails
- Evaluate the performance of the classifier
- Classify new emails as spam or not spam
- User-friendly GUI for easy interaction using PyQt5
from sklearn.feature_extraction.text import CountVectorizer
CountVectorizer is a class from the sklearn.feature_extraction.text module in the scikit-learn library. It is used to convert a collection of text documents to a matrix of token counts. This is a crucial step in text preprocessing for machine learning models, as it transforms text data into numerical data that can be used by algorithms. Each row in the resulting matrix represents a document, and each column represents a unique word (token) from the entire corpus. The value in each cell indicates the count of the word in the corresponding document.
from sklearn.naive_bayes import MultinomialNB
Multinomial Naive Bayes (MultinomialNB) is a variant of the Naive Bayes algorithm that is particularly suited for classification with discrete features (e.g., word counts for text classification). It assumes that the features follow a multinomial distribution. This classifier is often used for document classification problems, where the frequency of each word is used as a feature for training.
Key characteristics:
- Suitable for discrete data.
- Commonly used in text classification and natural language processing (NLP).
- Assumes that the features are conditionally independent given the class.
Parameters:
alpha: Additive (Laplace/Lidstone) smoothing parameter (0 for no smoothing).
Methods:
fit(X, y): Fit the model according to the given training data.predict(X): Perform classification on an array of test vectors X.predict_proba(X): Return probability estimates for the test vector X.score(X, y): Returns the mean accuracy on the given test data and labels.
To install the required dependencies, run:
pip install -r requirements.txtTo start the application, run:
python app.pyThis project is licensed under the MIT License.