How victorsepulveda1994/minor-machine-learning works
Project Components
• Core Pipeline (Proyecto_Final): This is organized into several structured stages:
• Etapa01-Obtain: Data ingestion—retrieving and loading raw human activity data.
• Etapa02-Scrub: Data cleaning and preprocessing—removing inconsistencies.
• Etapa03-Explore: Exploratory Data Analysis (EDA)—visualizing patterns and trends.
• Etapa04-Model: Training and evaluating models for human activity recognition.
• Etapa05-Interpret: Model interpretation and results explanation.
• Additional Learning Modules and Experiments:
• S07-Linear-Models contains a notebook focused on linear regression techniques.
• Tarea-Algebra appears to be an algebra-related exercise.
• Validación Cruzada includes a notebook and dataset for cross-validation, using the penguins dataset.
• The notebooks directory holds various exploratory notebooks on topics such as Support Vector Machines, transitioning from machine learning to deep learning, PyTorch usage, and an adaptation of OpenAI Whisper.
Out Architectural Relationships
-> Data Collection (Etapa01-Obtain) → Data Cleaning (Etapa02-Scrub) → EDA (Etapa03-Explore) → Modeling (Etapa04-Model) → Interpretation (Etapa05-Interpret)
• Experimental and Educational Paths:
• A branch denoting separate experimental notebooks (S07-Linear-Models, Validación Cruzada, and Tarea-Algebra) which, while not part of the main pipeline, represent complementary learning exercises.
• The general notebooks directory can be visualized as a side branch covering advanced topics (SVMs, PyTorch, deep learning strategies) that supplement the central narrative.
Architectural Patterns and Design Principles
• Pipeline-based Structure: The core structure of Proyecto_Final follows the archetypal ML workflow—a chain of distinct yet interconnected stages.
• Modular Organization: Each “Etapa” (stage) is modular and independent, making it easier to update, extend, or reuse components.
• Separation of Concerns: The experimental notebooks (e.g., algebra tasks, regression models, cross-validation) are maintained separately, emphasizing the educational purpose and allowing focused exploration of individual topics.
Relevant Technologies
Although explicit details on libraries or frameworks aren’t provided in the README, the content (Jupyter notebooks on diverse ML topics) implies usage of Python-based data science stacks such as:
• Data Processing/EDA: pandas, numpy, matplotlib, seaborn.
• Machine Learning/Deep Learning Frameworks: scikit-learn, PyTorch.
• Signal Processing/Specialized Models: Possibly libraries related to human activity recognition.
• Annotations as necessary to reflect the primary technology or tool being employed, even if abstractly inferred (e.g., “Jupyter Notebook-based analysis”).