Artificial Intelligence with Crypto Stocks

Learn how Bitcoin mining works as a decentralized computation that confirms transactions, issues new bitcoins, and uses proof of work with hashes, blocks, nonce, and mining difficulty.

Learn online effectively with active industry teachers, affordable, rewatchable content, and project-based learning that builds your portfolio through hands-on practice and real skills.

Machine learning is an algorithm that learns from data to automatically improve with experience. It includes supervised, unsupervised, and reinforcement learning and follows a workflow from data gathering to deployment.

Supervised learning trains a model to map inputs to outputs by inferring a function from training data and using it to predict new outcomes.

Explore building two machine learning models to predict cryptocurrency prices using blockchain features pulled from a free API, processed with Python, and visualized alongside linear and polynomial regression.

Learn linear regression, a simple supervised learning method that predicts a continuous value from features by fitting a linear relationship using least squares, even with noise or sparse data.

Collect data from blockchain API and blockchain.com charts to build features for predicting bitcoin price, export csvs, align dates, and join data with Python pandas.

Merge multiple CSV files in Google Colab using Python and pandas, building a single data frame with time and features like hash rate and transaction fees to model market price.

Remove the first row of the data frame to discard headers, drop the time column with axis one, and normalize and scale values for visualization in the next lecture.

Visualize data by plotting time on the x-axis against multiple features, after converting time to datetime and other columns to numeric, then build multiple charts to reveal feature trends.

Build a linear regression model with scikit-learn, train on x and y data, and visualize the best-fit line for bitcoin price versus transactions, then preview polynomial regression for curved relationships.

Build a polynomial regression model by adding higher-degree x terms via scikit learn's polynomial features, and compare degrees to identify the optimal fit.

Preview a project to cluster cryptocurrencies with unsupervised learning. Use k-means with elbow curve to determine the number of clusters, leveraging data from the cryptocompare api and pca dimensionality reduction.

Learn unsupervised learning and its applications in recommendation systems, customer grouping, and handwriting or speech recognition, plus its two main techniques: principal component analysis and cluster analysis.

Fetch crypto data from the Cryptocompare API in Google Colab, convert the data into a pandas data frame, and save it as a CSV to support an unsupervised clustering project.

Clean and prune the crypto data frame by dropping irrelevant columns and preserving the algorithm and proof of type to prepare for clustering.

Create dummy variables for text features with pandas get_dummies and combine them with the data. Scale the data and apply principal component analysis to reduce dimensionality for machine learning.

Explore principal component analysis as an unsupervised dimensionality reduction method, reducing high dimensionality by filtering noise, preventing overfitting and improving generalization for machine learning models.

Scale data with a standard scaler and reduce dimensionality using principal component analysis to three components, then prepare a PCA data frame for k-means clustering on crypto stocks.

Explore the theory of k-means clustering, an unsupervised learning technique that groups data into k clusters by iteratively updating centroids and assigning samples to the closest centroid.

Apply k-means clustering to group cryptocurrencies using the elbow method to determine the optimal number of clusters, leveraging PCA-transformed data and inertia to evaluate each k.

Compute the optimal four clusters from the elbow curve, train a k-means model on the SPCA data frame, and predict the cluster labels.

Visualize final results by plotting a 3D cluster graph of cryptocurrencies using Plotly Express, color- and symbol-coded by cluster, with principal components as axes.

Learn k nearest neighbors as a non-parametric, instance-based method for classification and regression, and explore how the k value affects boundaries and cross-validation error.

Build a k nearest neighbors classifier to predict crypto price direction using data scraped from the Yahoo Finance API, fetching Dogecoin (DOGE-USD) data in Colab and saving as CSV.

Learn to build a k nearest neighbors classifier for crypto prices by creating a price increase/decrease label and using high, low, and close as features, with data splitting and scaling.

Learn how radius neighbors regression predicts cryptocurrency stock price movements, using all samples within a radius and contrasting with k nearest neighbors, guided by density, normalization, and dimensionality.

Build and evaluate a radius neighbors regressor in scikit-learn to predict stock returns, train with your data, and explore radius values to improve model score.

Explore gradient boosting and its ensemble of weak models, mainly decision trees, for classification and regression. See how Catboost applies this method to crypto trading decisions and rankings.

Load XRP USD data via Yahoo Finance API in Google Colab, download data for 2014-01-01 to 2021-01-01, create a next-day buy/sell label, and train a Catboost model.

Build an XGBoost regressor to predict crypto prices, set squared error objective, and fit on training data; impute missing values and evaluate with mean absolute error.