Machine Learning Model Cheatsheet
In BigQuery ML, you can use a model with data from multiple BigQuery Datasets for training and for prediction.
| Task | Examples | ML Models |
|---|---|---|
| Predict values | 1. Predict sales figures 2. Predict stock prices | 1. Linear Regression 2. Boosted Trees Regressor 3. AutoML Table Regressor 4. DNN Regressor 5. Wide & Deep Regressor |
| Predict between categories | 1. Identify spam emails 2. Identify tumour types | 1. Logistic Regression 2. Boosted Trees Classifier 3. AutoML Table Classifier 4. DNN Classifier 5. Wide & Deep Classifier |
| Generate recommendations | 1. Product recommendations 2. Create personalised content | 1. Wide & Deep Classifier 2. Matrix Factorisation |
| Reduce data dimensionality | 1. Analysis of written text 2. Analysis of DNA data | 1. PCA 2. Autoencoder |
| Find anomalies | 1. Detect fraud 2. Predict credit risk | 1. PCA 2. Autoencoder 3. K-Means 4. ARIMA-PLUS |
| Group data into clusters | 1. Perform customer segmentation | 1. K-Means |
| Time series forecasting | 1. Predict housing prices based on historical data | 1. ARIMA-PLUS |
Post Tags: