Brand Name Generator

A sophisticated Python-based brand name generator that uses Markov chains and customizable weights to create unique, memorable brand names.

Features

• Markov Chain Generation: Creates natural-sounding names using configurable n-gram order
• Weighted Training: Support for multiple training corpora with adjustable weights
• Temperature Control: Adjustable randomness in name generation
• Letter Penalties: Customizable penalties for specific characters
• Length Control: Configurable minimum and maximum word lengths
• Case Handling: Automatic capitalization of generated names

Installation

git clone https://github.com/yourusername/brand-name-generator
cd brand-name-generator
python venv venv # optional
source venv/bin/acrivate # optional
pip install -r requirements.txt

Usage

Basic usage example:

from brand_name_generator import BrandNameGenerator

# Initialize generator
generator = BrandNameGenerator(order=2)

# Train with data
generator.train(["Example", "Words"], weight=1.0)

# Generate names
print(generator.generate_word(temperature=0.5))

Configuration

General Settings

• order: Length of character sequences used in generation (default: 2)
• min_length: Minimum length of generated names (default: 3)
• max_length: Maximum length of generated names (default: 12)
• temperature: Controls randomness (0.0-1.0, lower = more conservative)

Corpus Configuration

Train the model with different text corpora:

# General vocabulary
generator.train(general_corpus(), weight=1.0)

# Brand names (weighted higher)
generator.train(brand_names, weight=10000)

Letter Penalties

Customize character probabilities using the letter_penalties dictionary:

generator.letter_penalties = {
    'ä': 0.2,  # Reduce frequency of 'ä'
    'x': 0.3,  # Reduce frequency of 'x'
    'r': 2.0   # Increase frequency of 'r'
}

Default penalties are configured for optimal brand name generation in English/Finnish.

Advanced Usage

Custom Corpus Integration

# Load custom corpus
df = pd.read_csv('your_corpus.csv', sep='\t')
words = df["column_name"]
generator.train(words, weight=1.5)

Multiple Training Sources

# Layer different training sources with weights
generator.train(technical_terms, weight=0.5)
generator.train(existing_brands, weight=2.0)
generator.train(common_words, weight=1.0)

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License

Other useful tools

• https://www.domainnamesoup.com/six-letter-random-domain-names.php