Principles followed

Search Algorithm
Used in the FindIt Interface

General Principles followed

Consistency with native interface: The search of each source should match the native interface, so that the link "View more hits" yields the same number of hits as what is announced in the FindIt result page.
Usefulness of the results: If we can make certain adjustments while respecting Principle #1, we should strive to get back results from each source that are as useful as possible, with a good balance between recall and precision.

Resources searched by FindIt

FindIt searches across multiple Library resources:

E-Resource Locator, for identifying e-journals, databases, onlinedictionaries,etc. ( this includes the Research Guides)

Communities of interest and resource types

Franklin, the Penn Library online catalog

Journals, newspapers, magazines

Vcat, the Penn Library video catalog

Library A to Z web pages, including hours, borrowing information, Library floor plans

Penn Library FAQ

Penn Library Staff Directory

Search Algorithm for "Anywhere" Option

Source	Search Type	Fields Searched	Right Truncation	Stop words	Initial articles
E-Resources	Keywords, AND	Title, aliases, keywords, description	Yes, at the word level	Removed	Removed (as stop words), for English only
Communities and Resource Types	Keywords, AND Reposted as OR if AND fails	Name and aliases	No	Removed	Removed (as stop words), for English only
Franklin	Keywords, AND	Any keyword searchable fields	No	Removed	Removed (as stop words), for English only
Journals, newspapers, magazines (really a Franklin search)	Title search	Title field (limited on periodicals [bib_format='as'])	Yes, at the phrase level	Kept	Removed (for English, French, German, Italian, and Spanish articles)
Vcat	Keywords, AND	Any keyword searchable fields	Yes, at the word level	Removed	Removed (as stop words), for English only
Staff	Keywords, AND	Person's name, title, library name, library department, building, email address	Yes, at the word level	Removed	Removed (as stop words), for English only
A2Z	Keywords, AND	AtoZ Title, aliases, summary, keywords	Yes, at the word level	Removed	Removed (as stop words), for English only
FAQ	Keywords, AND	Question, keywords	Yes, at the word level	Removed	Removed (as stop words), for English only

Search Algorithm for "In E-Resource Title Only" Option

Source	Search Type	Fields Searched	Right Truncation	Stop words	Initial articles
E-Resources	Keywords, AND	Title and aliases	Yes, at the word level	Removed	Removed (as stop words), for English only

Relevancy algorithm

For each source or sub-category within a source, FindIt bubbles up any title that contains keywords from the query. The more keywords are presents in a title, the higher that title bubbles up.

So for instance if the search on the query “dogs and cats” brings back:

The canine journal
The journal of dogs
The journal of dogs and cats
The journal of pets

The list will be reordered to: