How do Metric Score Distributions affect the Type I Error Rate of Statistical Significance Tests in Information Retrieval?
Julián UrbanoMatteo CorsiAlan HanjalicPublished in: ICTIR (2021)
Keyphrases
- error rate
- significance tests
- score distributions
- information retrieval
- meta search
- relevant documents
- test set
- search engine
- information retrieval systems
- information filtering
- mixture model
- query expansion
- test collection
- expected loss
- text mining
- normal distribution
- hypothesis tests
- rank aggregation
- latent semantic indexing
- maximum likelihood
- relevance feedback
- retrieval effectiveness
- retrieval model
- document retrieval
- language model
- ranked list
- text retrieval
- image retrieval
- retrieval systems
- user queries