SMS Spam Collection

SMS Spam
Dataset

The SMS spam dataset from the UCI Machine Learning Repository contains 5,574 labeled messages for text classification and natural language processing tasks, making it a classic small dataset for NLP beginners.

Get the Dataset Now
5,574 messages 2 categories CC BY 4.0 license T. Almeida et al.
SMS Spam Collection
📊
5,574
Total Messages
📨
747
Spam Messages
✅
4,827
Normal Messages
📜
CC BY 4.0
Open License Agreement

Dataset Highlights

Classic NLP text classification dataset, suitable for beginners in natural language processing

📱

Real SMS Data

The data comes from real mobile text messages, including daily conversations and various spam promotional messages, with a natural and authentic language style.

🏷️

Binary Classification Labels

Each message is labeled as ham (normal) or spam (junk), with high labeling quality, suitable for supervised learning.

📝

NLP Beginner Friendly

Text length is moderate, and language is concise, making it very suitable for beginners to learn text preprocessing, TF-IDF, and bag-of-words models.

⚖️

Class Imbalance

Spam messages account for only 13.4%, allowing practice with various techniques for handling class imbalance issues.

🔧

Applicable to Multiple Methods

Can be used for comparative experiments of various text classification methods such as Naive Bayes, SVM, and deep learning.

🏛️

UCI Authoritative Source

Originating from the UCI Machine Learning Repository, widely used as a standard text classification benchmark in the NLP community.

Applicable Scenarios

From basic text classification to advanced NLP, with a rich variety of application scenarios

📧

Spam Detection

Build a spam text message filter, the most classic application scenario for Naive Bayes classification

📝

Text Preprocessing

Practice techniques for tokenization, removing stop words, TF-IDF, and other text feature extraction methods

🤖

Deep Learning

Use models like LSTM and BERT for text classification, comparing with traditional methods

⚖️

Imbalance Handling

Practice techniques for handling class imbalance such as oversampling, undersampling, and weighted loss

Natural Language Processing Text Classification Spam Detection Beginner Dataset Naive Bayes

Data Preview

The following are examples of the first few rows of the SMS spam dataset

CSV
label,message
ham,"Go until jurong point, crazy.. Available only in bugis n great world la e buffet..."
ham,Ok lar... Joking wif u oni...
spam,"Free entry in 2 a wkly comp to win FA Cup final tkts 21st May 2005."
ham,"U dun say so early hor... U c already then say..."
ham,"Nah I don't think he goes to usf, he lives around here though"

3 Steps to Get Started Quickly

From browsing to analysis, you can start your data science project in just a few minutes

01

Browse the Dataset

View dataset details on the Ace Data Cloud platform, understand field descriptions, sample size, and licensing agreements.

02

Download Data

Download the CSV file (486 KB), which contains all labeled SMS data.

03

Load and Analyze

Use pandas.read_csv() to load the data, along with sklearn.feature_extraction.text for feature extraction.

Get the Dataset

Start Exploring SMS Classification Data

A classic NLP dataset, open license, available for immediate download. 5,574 real labeled SMS messages, ideal for getting started with text classification and spam detection.

Get the Dataset Now