Portfolio Artifact • Applied Machine Learning

Machine Learning Algorithms: Visual Framework

This portfolio artifact presents a structured infographic of major machine learning algorithms and shows how they relate across supervised learning, unsupervised learning, and deep learning. It was designed to demonstrate conceptual understanding, technical communication, and professional portfolio presentation for applied AI leadership and interdisciplinary collaboration.

Professional Framing

As an emerging AI leader with a cybersecurity and applied systems perspective, I created this artifact to translate complex machine learning concepts into a format that is visually clear, technically accurate, and accessible to both technical and non-technical audiences.

The framework emphasizes how algorithm choice depends on problem type, data structure, and application domain. Rather than listing models in isolation, it presents them as part of a connected machine learning landscape.

AI Leadership Technical Communication Portfolio Artifact Academic Showcase

Artifact Snapshot

Focus: Machine learning taxonomy and practical application

Coverage: 10 major algorithms across 4 AI domains

Domains: Tabular Data, Computer Vision, NLP, Generative AI

Purpose: Show relationships, differences, and real-world use

Supervised Learning Unsupervised Learning Deep Learning / Modern AI 📊 Tabular Data 🖼️ Computer Vision 💬 NLP ✨ Generative AI

Artifact Overview

Title Machine Learning Algorithms: Visual Framework

Introduction This artifact provides a visual overview of core machine learning algorithms and organizes them by learning style, domain relevance, and real-world application.

Objective To classify major algorithms clearly, explain how they work, and demonstrate how they support different AI problem spaces.

Process I researched representative algorithms, grouped them into supervised, unsupervised, and deep learning categories, and designed a visual structure that communicates both distinctions and relationships.

Professional Value

Tools and Technologies Used HTML, CSS, GitHub Pages, machine learning concepts, portfolio design principles

Value Proposition Demonstrates my ability to synthesize technical material into a polished artifact for academic and professional audiences.

Unique Value Combines technical understanding, visual communication, and strategic AI framing in one portfolio-ready deliverable.

Relevance Supports my long-term goal of contributing to and leading AI initiatives that require both analytical understanding and strong communication.

From General Categories to Specific Algorithms

Machine Learning

Systems that learn patterns from data in order to predict, classify, group, or generate outputs.

Supervised Learning

Uses labeled data to learn direct relationships between inputs and known outputs.

Linear Regression

Supervised 📊 Tabular Data

How it works: Fits a line that models the relationship between input variables and a numeric outcome.

Use case: Predicting house prices or forecasting revenue.

Logistic Regression

Supervised 📊 Tabular Data 💬 NLP

How it works: Estimates the probability that an input belongs to a category.

Use case: Spam detection or customer churn classification.

Decision Tree

Supervised 📊 Tabular Data

How it works: Splits data into branches using sequential if-then decisions until a final class or value is reached.

Use case: Credit risk analysis or fruit classification.

Random Forest

Supervised 📊 Tabular Data

How it works: Combines multiple decision trees and uses their collective vote for more stable predictions.

Use case: Fraud detection and customer behavior analysis.

Support Vector Machine

Supervised 📊 Tabular Data 🖼️ Computer Vision

How it works: Finds the best boundary that separates classes with the widest margin.

Use case: Image categorization and diagnostic support.

Unsupervised Learning

Works with unlabeled data to discover hidden patterns, natural groups, or simplified structure.

K-Means Clustering

Unsupervised 📊 Tabular Data

How it works: Groups data points into clusters around central averages called centroids.

Use case: Customer segmentation and product grouping.

Hierarchical Clustering

Unsupervised 📊 Tabular Data

How it works: Builds a layered tree of clusters by merging or splitting similar data points.

Use case: Similarity analysis and gene grouping.

Principal Component Analysis (PCA)

Unsupervised 📊 Tabular Data 🖼️ Computer Vision

How it works: Reduces many features into a smaller set of components while preserving important variation.

Use case: Dimensionality reduction and data visualization.

Deep Learning / Modern AI

Uses layered neural networks to learn complex representations from unstructured and high-volume data.

Convolutional Neural Network (CNN)

Deep Learning 🖼️ Computer Vision

How it works: Learns visual features by scanning images for shapes, textures, and spatial patterns.

Use case: Facial recognition, medical imaging, and object detection.

Recurrent Neural Network (RNN / LSTM)

Deep Learning 💬 NLP

How it works: Processes sequence data step by step while retaining information from earlier inputs.

Use case: Time-series forecasting and earlier language modeling tasks.

Transformer

Deep Learning 💬 NLP ✨ Generative AI

How it works: Uses attention mechanisms to model relationships between all parts of the input at once.

Use case: ChatGPT, summarization, translation, and text generation.

How the Algorithms Relate

Supervised algorithms are best when labeled outcomes are available and direct prediction is required.
Unsupervised algorithms help identify patterns without labels, making them valuable for segmentation and structure discovery.
Deep learning models extend machine learning into complex domains such as images, language, and generative systems.
Classical models are often more interpretable, while deep learning is typically stronger for unstructured data.

Application Domains

📊 Tabular Data Used for forecasting, fraud detection, customer analytics, and structured decision support.

🖼️ Computer Vision Used for image classification, facial recognition, object detection, and medical image review.

💬 NLP Used for sentiment analysis, translation, speech recognition, and question answering.

✨ Generative AI Used for chatbots, text generation, content creation, and multimodal AI assistants.

Reflection on Learning

Creating this artifact strengthened my understanding of how machine learning algorithms connect across different problem types and data domains. Instead of viewing each model as an isolated concept, I learned to organize them within a broader analytical framework that distinguishes supervised, unsupervised, and deep learning approaches.

This project also reinforced the importance of choosing algorithms based on both the structure of the data and the intended outcome. In addition, it improved my ability to communicate technical concepts in a visually clear and professionally structured format, which is a critical skill for collaboration and leadership in AI-focused roles.

Key Takeaways

Algorithm choice should match the data type and the business objective.
Supervised learning is strongest for labeled prediction tasks.
Unsupervised learning is useful for identifying hidden patterns and groups.
Deep learning expands AI capability in language, vision, and generative systems.
Clear communication of technical ideas is part of effective AI leadership.