User Manual: Search

• Content Words and Phrases
• Fuzzy Searches
• Wild Card Searches
• Boolean Operators
• Positional Operators: Context Related
• Precedence and Parentheses
• Number Range Operator
• Quorum Operator
• Separators
• Search Rules and Conventions

Search Capabilities

GovernmentDocs.org provides enterprise search capabilities to users who want to perform sophisticated context-specific searches to find targeted results within the large set of government documents.   The search is driven by optical character recognition (OCR) technology, which translates digital images of text to machine-editable text that is stored using Extensible Markup Language (XML).  All documents uploaded to GovernmentDocs.org are OCRed and the resulting text is then indexed for retrieval.

GovernmentDocs.org’s search engine, powered by ZyLab, was developed specifically for use by law firms for e-discovery and to assist government record management.  E-discovery refers to the discovery of electronically stored information for use as evidence in a civil or criminal legal case. Therefore, GovernmentDocs.org is powered by a sophisticated search engine and has numerous techniques for customizing searches.  These techniques and the associated syntax are discussed in Section II of this manual.

There are two functional areas that also provide end-users with a significant advantage when trying to find specific information:

• Relevance and Ranking
• Hit Highlighting and Hit Navigation

Relevance and Ranking i

When search engines index the Internet, they typically only want to find the best sites, or the most popular ones, as opposed to finding all potentially relevant sites. Thus search platforms that are derived from these Internet search engines focus on 100% precision and use deeply embedded relevance ranking algorithms, which can be based upon, for instance, popularity ranking (in traffic and in links).

In many legal, intelligence or investigative environments, though, “popular” hits aren’t always desired; for example, consider the complex financial transactions in a money laundering case. Users in this instance cannot miss a document that contains a “smoking gun” or provides critical details that could influence the case. All potentially relevant hits must be found, meaning 100% recall is needed, which is difficult to achieve if an engine is optimized for precision across an implementation.  GovernmentDocs.org’s search engine returns all relevant results and then arranges the set of retrieved records so that those most likely to be relevant to your request are shown to you first. There are a number of different factors that GovernmentDocs.org uses to determine relevance, such as density, frequency and breadth of match.

Hit Highlighting and Hit Navigation

Hit highlighting is defined as the search engines ability to highlight using a yellow color (similar to a yellow highlighter) where each occurrence of search terms is on a page. This feature can be used as a quick mechanism to locate the terms and determine whether a document is relevant or not. This is an important and essential feature of GovernmentDocs.org’s search engine.

It is also extremely important to be able to easily move between pages within the search results set. Without returning to the search results page, GovernmentDocs.org provides an easy mechanism to navigate to the next, or previous, page returned as part of the end user’s search.

II. Search Techniques and Syntax ii

Please review this section to learn more about the techniques and syntax that can be used to control the results provided by Governmentdocs.org’s sophisticated search engine. Almost any kind of search can be carried out. From explorative searches that will provide a broad sense of the available information to very precise searches.

Use Fuzzy search to find all occurrences of a word, including the ones that were not recognized correctly by the OCR engine. You can choose from 4 fuzzy degrees (the degree of closeness to your query). Fuzzy degree 2 is recommended for normal text searches. This provides for mistakes such as broken and joined characters. Set Fuzzy degree 3 or 4, if you search for long words.

The retrieval of irrelevant shorter words is prevented. For example, if the Fuzzy Degree is 4 and the search term is six characters long, the actual Degree of Fuzzy will be 0.5 X 6 = 3 and not 4.

• ? replaces one character.
• * replaces zero or more characters.

Boolean operators define the relationships between words or groups of words.

• AND
• OR
• NOT

Use positional operators to ensure that your search terms are contextually related. This is especially useful when searching long documents.

When you use two or more operators, the order in which these operators are executed is determined by operator precedence. You can override this order by using parentheses (grouped terms). Content within parentheses is interpreted as one unit. Always use parentheses in complex search statements (more than two operators).

The following operator precedence is executed:

• NOT
• OR
• W/n, P/n (these operators are of equal precedence)
• AND
• TO

You can use these math operators in number range searches:

• < less than
• < = less than or equal to
• = equal to
• < > not equal to
• > greater than
• > = greater than or equal to

Use Quorum operators to search for a specified number of terms within a search statement.

n of {term, term, ...}

Use Separators to limit a search to a physically defined range of a text file. Separators are very useful when combined with the TO operator.

You can use these separators:

• EOP end of paragraph
• EOG end of page
• EOS end of sentence
• EOL end of line
• EOG end of page

• Operators are placed only between search terms, except NOT.
• NOT can be placed at the beginning of a search statement, but never at the end.
• Two operators cannot appear in sequence, except NOT (AND NOT, OR NOT).
• Operators are noise words, therefore, you cannot use them as content words. The search statement and OR or will not be accepted.
• Uppercase and lowercase letters are considered identical. We show operators in upper case for emphasis and clarity.
• An operator can appear more than once in a search statement.
• The W/n operator must include an integer in the range 1 to 16382, followed by a space and a content word. 
• You can use one term to retrieve both the hyphenated and non-hyphenated spellings of a term; for example, the search term:
• database retrieves database and data-base, but not data base
• data-base retrieves data-base, but not database and data base
• data base retrieves data-base and data base, but not database.
• Searches with duplicate words in series also find single occurrences of that word; for example, the search statement, sing sing, would find single occurrences of sing as well as the phrase, sing sing.
• Words with normally appearing hyphens are recognized, for example, Winston-Salem.
• All printable characters in the ASCII character set are recognized.
• A sentence-ending period and other trailing punctuation marks are ignored, when a space or a carriage return follows. Periods are recognized when followed by a character, as in I.B.M. or in 292.004. Apostrophes are treated as null characters, and ignored.