[SPARK-10413] ML models should support prediction on single instances

• Type: Umbrella
• Status: Resolved
• Priority: Critical
• Resolution: Incomplete
• Component/s: ML
• Labels: bulk-closed
• Assignee: Unassigned
• Reporter: Xiangrui Meng
• Created: 02/Sep/15 05:16
• Updated: 26/Dec/19 11:19
• Resolved: 08/Oct/19 05:44

Description

Currently models in the pipeline API only implement transform(DataFrame). It would be quite useful to support prediction on single instance.

UPDATE: This issue is for making predictions with single models. We can make methods like def predict(features: Vector): Double public.

• This issue is not for single-instance prediction for full Pipelines, which would require making predictions on {{Row}}s.

Issue Links

• Is contained by: SPARK-14377 Review spark.ml parity for classification, except trees
• is related to: SPARK-30358 ML expose predictRaw and predictProbability
• is related to: SPARK-16365 Ideas for moving "mllib-local" forward
• relates to: SPARK-16431 Add a unified method that accepts single instances to feature transformers and predictors
• relates to: SPARK-5874 How to improve the current ML pipeline API?

Sub-Tasks

• Support prediction on single instance for regression and classification related models: Resolved
• Support prediction on single instance in KMeans/BiKMeans/GMM: Resolved