pkmb 0.1.2

A comprehensive NLP and Machine Learning package with example implementations

PKMB Package

A comprehensive Python package containing various NLP and Machine Learning implementations.

Installation

pip install pkmb

Usage

from pkmb import print_program

# Print any program (1-5, 7, 9)
print_program(1)  # Basic NLP operations
print_program(2)  # Named Entity Recognition
print_program(3)  # TF-IDF implementation
print_program(4)  # N-grams analysis
print_program(5)  # Word Embeddings analysis
print_program(7)  # Text Generation with LSTM
print_program(9)  # Variational Autoencoder for MNIST

Available Programs

1. Program 1: Natural Language Processing (NLP) Text Analysis

   • Basic NLP operations using NLTK
   • Includes: tokenization, stopword removal, stemming, and lemmatization
   • Demonstrates both sentence and word-level processing

2. Program 2: Named Entity Recognition (NER)

   • Uses NLTK for entity extraction
   • Identifies persons, organizations, locations
   • Includes BIO tagging and tree representation

3. Program 3: TF-IDF Implementation

   • Manual implementation of TF-IDF calculation
   • Comparison with scikit-learn's TfidfVectorizer
   • Document similarity analysis

4. Program 4: N-grams Analysis

   • Uses Pride and Prejudice as corpus
   • Generates unigrams, bigrams, and trigrams
   • Includes frequency analysis and visualization

5. Program 5: Word Embeddings Analysis

   • Uses GloVe embeddings (50d)
   • Word similarity computation
   • Semantic relationship analysis

6. Program 7: Text Generation with LSTM

   • Neural network-based text generation
   • Uses TensorFlow/Keras LSTM architecture
   • Includes training and text generation capabilities

7. Program 9: Variational Autoencoder (VAE)

   • Deep learning model for MNIST dataset
   • Implements both encoder and decoder networks
   • Generates new digit images from latent space

Note: Programs 6 and 8 are intentionally omitted from this collection.

Dependencies

The package requires the following Python packages:

pip install nltk pandas scikit-learn requests gensim scipy==1.11.4 tensorflow matplotlib numpy

Additional Setup

1. NLTK Data: Required for Programs 1-4

   • Downloads automatically when running the programs
   • Includes: punkt, stopwords, wordnet, averaged_perceptron_tagger, maxent_ne_chunker

2. GloVe Embeddings: Required for Program 5

   • Downloads automatically on first use (~66MB)
   • Uses the glove-wiki-gigaword-50 model

3. MNIST Dataset: Required for Program 9

   • Downloads automatically through TensorFlow
   • Used for training and testing the VAE

Note on GPU Support

Programs 7 (LSTM) and 9 (VAE) can benefit from GPU acceleration if TensorFlow is installed with CUDA support.

Error Handling

All programs include proper error handling and will display informative messages if:

• Required data is not available
• Words are not found in vocabulary
• Models fail to load or process