How saigunaranjan/probability_and_machinelearning works
Project Type and Purpose
• This repo is a collection of standalone Python scripts and modules for probability-theory experiments and machine-learning exercises.
• There is no web front-end or microservice layer—rather it’s a modular, script-based toolkit.
File Structure
Top-level items:
• binary_classification.py, support_vector_machine.py, expectation_maximization_algo.py, buffonsNeedle.py, parabolicConstant.py
• ml_functions_lib.py (shared utility library)
• datasets/ (CSV files, e.g. Iris.csv)
• deep_neural_networks/ (five specialized NN scripts)
Notes:
– Each script imports from ml_functions_lib.py or directly loads datasets/*.
– No deployment scripts—these run locally as Python programs.
Extract Insights from the README
• The README confirms this is a “Probability_and_MachineLearning” toolkit for modelling and exercises.
Main Components & Their Relationships
a. Components:
1. Data Storage
– datasets/Iris.csv (and any CSVs you may add)
2. Shared Utilities
– ml_functions_lib.py
3. Probability‐Theory Scripts
– buffonsNeedle.py
– parabolicConstant.py
– expectation_maximization_algo.py
4. Classical ML Scripts
– binary_classification.py
– support_vector_machine.py
5. Deep Learning Suite
– deep_neural_networks/ (xor_gate.py, iris_data.py, multilayer_feedforward_nn.py, digit_recognition.py, disease_classification.py)
b. Interactions:
– All ML and probability scripts read from datasets/*.
– Scripts call into ml_functions_lib.py for common routines (e.g., matrix ops, activation functions).
– Deep neural scripts may also depend on external libs (NumPy, Pandas, TensorFlow/PyTorch) – annotate these as external services.
c. Architectural Patterns:
– Modular monolith: logically separated modules within one codebase.
– Reusable utility layer (ml_functions_lib) that prevents code duplication.
d. Technologies/Frameworks:
– Python 3.x
– Pandas/NumPy/Scikit-Learn/TensorFlow or PyTorch (as implied by deep_neural_networks)
Detailed Capturing of Architectural Elements
– Group scripts into three logical tiers: Probability, Classical ML, and Deep Learning.
– Place ml_functions_lib centrally as the shared utility layer.
– Annotate each group with bullets listing contained scripts.
– Note all external library dependencies off to one side to show environment requirements.
[Datasets (Iris.csv, …)] →
• buffonsNeedle.py
• parabolicConstant.py
• expectation_maximization_algo.py
• binary_classification.py
• support_vector_machine.py
• xor_gate.py
• iris_data.py
• multilayer_feedforward_nn.py
• digit_recognition.py
• disease_classification.py
→ [External Libraries: NumPy, Pandas, Scikit-Learn, TensorFlow/PyTorch]
– Blue for Data Store,
– Green for Probability,
– Yellow for ML,
– Orange for Deep Learning,
– Gray for Utilities,
• Highly detailed—every script/module is called out.
• Accurate—reflects actual code organization and dependencies.