table of Contents
Chapter 8 queries users in understanding and O2O scene guide
8.1 principles of modern search engine
8.2 accurate understanding of inquiry
Under Chapter 9 O2O scene sorting features
9.3 orthogonal multilayer A / B test
Chapter 10 in recommended application O2O scene
10.1 Typical recommended scene O2O
10.2 O2O recommend scene features
10.3 US group recommended practices - Recommended Framework
10.4 US group recommended practices - Recommended recall
10.5 US group recommended practices - Recommended sort
Chapter 8 queries users in understanding and O2O scene guide
8.1 principles of modern search engine
8.2 accurate understanding of inquiry
By understanding the need to consider the following query:
Intention recognition, entity recognition, speech tagging, query rewrite, tightness computing, word weight calculation, query correction
8.2.1 Definition and Identification of the user's query intent
How to define the user's query intent?
How to identify the user's query intent?
The user intent regarded as classification, clustering, topic distribution
8.2.2 Entity Recognition and Structured Query
8.2.3 Changes in the recall strategy
8.2.4 query rewrite
8.2.5 word weights and correlation calculation
8.2.6 Category correlation with the manual annotation
8.3 guides the user searches
Before the search guide (recommended), guided search (complement), after a search guide (Search)
Under Chapter 9 O2O scene sorting features
O2O sort of features:
US Mission ordering scheme:
9.1 Overview
9.2 Online ordering service
9.3 orthogonal multilayer A / B test
9.4 feature acquiring
Offline Survey System 9.5
Project 9.6 features
9.7 Ranking Model
9.8 sort of scene
Chapter 10 in recommended application O2O scene
10.1 Typical recommended scene O2O
Guess you like, details page recommendation, recommend nearby
10.2 O2O recommend scene features
Location, user behavior history, real-time recommendations
10.3 US group recommended practices - Recommended Framework
10.4 US group recommended practices - Recommended recall
Recall based collaborative filtering
Based on the location of the recall
Based on the recall of the search query
Based recall graph
Based on real-time user behavior recall
Substitute Strategy
10.5 US group recommended practices - Recommended sort
Sort feature
Ordering samples
Ranking Model
Recommended evaluation