Have you ever wondered if there was a way to automatically classify the spam emails that arrive every day?
In this article,How to make a "spam mail detector" using PythonWe will explain this in an easy-to-understand manner, even for beginners.
Using the machine learning library "scikit-learn," we analyze the content of emails using natural language processing.Program to automatically detect spamWe will implement the following.
You may think, "It seems difficult...", but don't worry, everything you need to know is explained in detail.
On this occasion,Let's create your own AI that automatically sorts spam emails!
- What is a spam detector? [Purpose and use cases]
- How to build a Python environment [for beginners]
- Data preparation and preprocessing [Natural Language Processing]
- Model training and evaluation [Implementation of machine learning]
- Trying out actual spam detection [Prediction]
- Summary | Ideas for future applications
What is a spam detector? [Purpose and use cases]
✅Conclusion
What is a spam detector?A program that analyzes the contents of emails and classifies them as spam or not.is.
✅Reason
In the internet society, we receive spam mails such as scams and advertisements on a daily basis. It is inefficient to sort them manually, soAutomatic discrimination by AIThere is a strong demand for this.
✅Specific examples
For example, it can be used for the following purposes:
• Integrate with email services for automatic sorting
• Implement it into your company's internal systems as a security measure
• Can also be applied to chats such as LINE and Slack
✅point
Like thisTechnology to "reduce human effort"As such, spam detection is widely used.
How to build a Python environment [for beginners]
✅Conclusion
Getting started with the spam detector is easy; just install Python and the necessary libraries.
✅How to do it
1. Install Python
Download the latest version from the official website (https://www.python.org/).
2. Create a virtual environment
python -m venv spamenv source spamenv/bin/activate # Mac/Linux spamenv\Scripts\activate # Windows
3. Installing the library
pip install pandas scikit-learn matplotlib
4. Prepare Jupyter Notebook (if necessary)
pip install jupyterlab
✅ Supplementary information
Jupyter makes it easy to check and execute code, and is recommended for beginners.
Data preparation and preprocessing [Natural Language Processing]
✅Conclusion
To determine spamNatural Language Processing (NLP) to process email textis required.
✅Things to do
• Loading a dataset (e.g. SMS Spam Collection Dataset)
• Removing unnecessary symbols and spaces
• Converting word frequency to a number (Bag of Words)
import pandas as pd from sklearn.feature_extraction.text import CountVectorizer df = pd.read_csv("spam.csv", encoding="latin-1")[["v1", "v2"]] df.columns = ["label", "message"] vectorizer = CountVectorizer() X = vectorizer.fit_transform(df["message"]) y = df["label"].map({"ham": 0, "spam": 1})
✅Points
Converting text into numbers using natural language processing is the first step in using machine learningIt will be.
Model training and evaluation [Implementation of machine learning]
✅Conclusion
scikit-learnDecision trees and Naive Bayesmakes learning and assessment easy.
from sklearn.model_selection import train_test_split from sklearn.naive_bayes import MultinomialNB from sklearn.metrics import accuracy_score print("Accuracy rate:", accuracy_score(y_test, y_pred))
✅Points
• It is important to separate training data from test datais.
• Check the accuracy rate (precision) of the prediction results to judge performance.
Trying out actual spam detection [Prediction]
✅Conclusion
Simply input the email text into the trained model,"Spam or not?"can be determined.
msg = ["Congratulations! You've won a prize!"] msg_vec = vectorizer.transform(msg) print(model.predict(msg_vec)) # → [1] (spam)
✅Points
• 1 is spam, 0 is good messageIt shows:
• You can see the effect clearly by trying it with multiple actual email texts.
Summary | Ideas for future applications
✅Key points
• Spam mail detection can be implemented using Python and scikit-learn
• The key is to convert text to numbers using NLP
• The model can achieve accuracy of 90% or higher.
✅Application examples
• LINE and SNS spam detection
• AI to monitor children online
• Automatic blocking tool for fraudulent emails
English 
Japanese