NPIC InfoBase Search Help


Contents:

Choosing Libraries
Choosing the Search Type
Conceptual Search
Boolean Search
Entering Search Terms
Entering Concept Search Terms
Advanced Conceptual Search Features
Controlling Conceptual Search Term Expansion
Entering Boolean Search Terms
Advanced Boolean Search Features
Executing the search
The Results List
Search Buttons
Changing the Result Sorting
Ranking (Conceptual Search only)
Viewing the Matching Document
Viewing a Summary of Hits
Refining results

Choosing Libraries

Libraries in the NPIC InfoBase are collections of documents related by topic or source. Choosing a Library chooses the source of documents your search will look through. Currently, the following Libraries are available:

EPA OPP E-Docket Document Archive, 2002-2008
This Library contains archived documents from the http://www.regulations.gov/ site. Please note that only EPA OPP (Office of Pesticide Programs) E-dockets are available.
Code of Federal Regulations Title 40
This Library contains the Code of Federal Regulations Title 40 - Pesticide Programs, Subchapter E
Electronic Code of Federal Regulations - Title 40
This Library contains a more-frequently-updated version of the Code of Federal Regulations Title 40 - Pesticide Programs, Subchapter E, from the new e-CFR site. Please note that the e-CFR site is not an official legal edition of the CFR, yet.

You must select at least one Library before executing the search. To facilitate selecting and deselecting multiple libraries, Select All and Select None buttons are provided.

Choosing several libraries will expand your search across each selected data source, but may slightly slow the result processing. Choosing fewer libraries will focus your search to the specific data source, and can be a bit faster.


Choosing the Search Type

The NPIC InfoBase utilizes technology that allows two powerful, but quite different, search mechanisms: Conceptual and Boolean search. Use the details below to decide which mechanism will work best for you. When in doubt, choose the Conceptual search mode, it's the easiest to use.

Conceptual Search

In a Conceptual search, you enter the query in "plain language", without the need to add any special search-control operators. The lack of special control terms makes Conceptual search the easiest to use of the two available types.

Each word you enter into the search box will be expanded into its synonyms, according to limits you may choose to set. The search is then performed using this expanded set of terms, so that relevant documents will be returned, even if the exact term you used doesn't appear in the document.

Concept searches work best when you are looking for general information on a topic, and would be interested in other, related topics that might not match the exact terms you are using.

Boolean Search

In a Boolean search, you enter words you would like to find, separated by control terms that specify how the words must appear in the document. Unlike conceptual search, no synonym expansion is performed, although simple variations like plurals will match.

Boolean searches can be both faster and more accurate than Concept searches, but require more effort on the part of the user. The Boolean terms that control the search can be used to build powerful search strings that will bring back a tightly-targeted set of matching documents.


Entering Search Terms

After choosing the type of search you would like to perform, simply enter words you would like to find and any optional control operators in the input boxes available. Specific details for Conceptual and Boolean search entry behavior are available in the sections below:

Entering Concept Search Terms

A single text entry field is provided for you to enter Concept search terms. Simply type the words you would like to find into this box. You may optionally add special behavior operators to adjust some aspects of the search process. A list of the operators is available below:

Advanced Conceptual Search Features

You can use any of the following advanced operators in Concept mode. Terms with these operators won't be expanded in the same way as normal conceptual search terms:

Tilde (~) for fuzzy matching
Place a tilde (~) as the first character of a search term to allow "fuzzy" pattern matching of that term. Use this if you aren't sure of the spelling of the term. As an example, a search for ~daizinon would return hits to the properly-spelled word diazinon. A button is available to allow you to select terms for fuzzy matching -- simply click the Unsure of spelling? button before performing your search to use this feature.
Double quotes ("") for an exact phrase
Placing a set of search terms between double quotes ("") will force the system to evaluate those terms as an exact phrase. This will ensure that matching documents will contain the words in the order shown, without other indexed words separating them. Non-indexed words (common words like "it", "a", "the" etc.) may still appear between the search terms, because the search system does not "see" certain common words as part of the document.
Parentheses ( ) for grouped terms
You can group terms together using parentheses, so that they'll be expanded in context of each other. While parentheses can be used in Concept mode, they are most useful in Boolean mode, see the advanced Boolean usage section for more information.
Wildcards (?, *, _, @, #, ^, [search expression])
You can use any of these characters to substitute a character or set of characters in a word. Wildcards can be used multiple times in one word. An explanation of each of the wildcards appears below.
Wildcard Usage Example
@ match exactly one alphabetical character p@st would match pest, past and post
# match exactly one numerical character 199# would match 1990 through 1999
* match zero or more characters pest* would match pest, pests, pesticide, etc.
_ match exactly one character plan_ would match plant, plans, plane, etc., but not plan
[ ] Match any single character in the brackets. A hyphen may be used to indicate a range of characters or numbers. past[ea] would match both paste and pasta
[^ ] Match any single character not in the brackets. A hyphen may be used to indicate a range of characters or numbers. 199[^8-9] would match 1990 through 1997

Controlling Conceptual Search Term Expansion

In a Conceptual search, you can control to what level your search terms are expanded. Use the Conceptual Expansion Level: menu box to choose the expansion level to use for your search.

Depending on the chosen level of expansion, a search for the word food might find similar words including feed, provision, victual, cuisine or many other related terms.

The available expansion levels are:

The default setting of "Exact Matches" is the most restrictive, and will offer the best speed and most limited set of matching documents. It's best to try this setting first, and if you receive too few matching documents, try increasing the expansion to a higher level.

Entering Boolean Search Terms

In Boolean mode, four text entry boxes are provided, along with buttons for choosing the most common Boolean control terms that will be used to join the terms entered in these boxes (see the table below for all available Boolean operators) . The simplest way to build a powerful Boolean search string is to enter terms you must have in the result into the first entry box. Then, use the other entry boxes to further specify your search by entering additional terms, and clicking the appropriate Boolean control term between the entry boxes.

If you fill all four boxes, or if you would simply like to see how the combined search string will look, click the Boolean button near the top of the screen. The server will parse your search string, add the appropriate Boolean terms to it in the specified order, and re-display it in the first text entry box. You can then modify the string directly, or add more search terms into the other three entry boxes and repeat the process.

The available Boolean operators that can be used are:

Operator Usage Description
AND word AND otherword both word and otherword must be found, anywhere in the document.
OR word OR otherword Either word or otherword must be found in the document
NOT word NOT otherword word must be found in the document, but the document can't contain otherword
WITHIN word otherword WITHIN N word and otherword must both be in the document, and otherword must occur within n words before or after word
ADJ word otherword ADJ N word and otherword must both be in the document, and otherword must come no more than n words after word
BETWEEN word BETWEEN otherword AND thirdword word, otherword, and thirdword must all be in the document, and word must occur between otherword and thirdword

If no Boolean term is found in a search string, the default is for AND to be used between terms.

Advanced Boolean Search Features

In addition to the Boolean operators described above, you may optionally use the same advanced operators as Concept mode described above. These additional operators can be used in conjunction with the Boolean terms to provide additional powerful search features. Of particular note is using parentheses around certain search terms to separate and group them -- use this feature to design targeted search strings for accurate results.

Once you have mastered the use of Boolean terms, phrases, and parentheses grouping, you will see how using these powerful features lets you find exactly the documents you want, without needing to wade through long lists of irrelevant results. Some examples of using Boolean search terms both with and without some advanced operators are shown below. You are encouraged to try these examples to see how results change depending on the way a search is phrased.

Search Example Will Result In
mosquito AND west AND nile Documents containing any of the terms, anywhere in the document. Note that "west" and "nile" need not appear next to each other for a document to match.
mosquito west nile Same as above -- if no Boolean terms are entered, the default is an implied "AND" between each term.
mosquito AND "west nile" Documents containing both "mosquito" and the phrase "west nile" -- by placing search terms in quotes, the search system will treat the terms as a phrase and only find matching documents that include those terms together.
mosquito AND west nile WITHIN 5 Documents containing the words "mosquito" and "west" with the word "nile" occurring within 5 words either before or after the word "west." Both the WITHIN and ADJ operators can be used if you want to find words near other words, but not necessarily an exact phrase. The next example shows the use of ADJ.
mosquito AND west nile ADJ 5 Similar to the previous search. ADJ ("adjacent") works like WITHIN, but forces the match to occur after the specified word, instead of on either side. In this case, it means our search would only match documents where "nile" occurred 5 or fewer words after the word "west."
mosquito AND "west nile" NOT crow Documents containing both "mosquito" and the phrase "west nile", but that do not contain the word "crow"
mosquito AND "west nile" NOT crow OR bird Most likely not what is expected. This search will result in any document that contains both "mosquito," and "west nile" or just "bird," but also does not contain "crow". See the next entry for an example of what would provide the likely intended result.
mosquito AND "west nile" NOT (crow OR bird) This search will result in any document that contains "mosquito" and "west nile", but not "bird" or "crow". The use of parentheses around terms that should be grouped together forces the Boolean control terms to apply to the desired search words, not just the nearest one.
(herbicide OR "weed killer") AND
((safe OR "low risk") NOT (danger OR caution))
This search illustrates a more complicated use of term grouping, including "nesting" of grouped search terms inside another group. Along with matching either the word "herbicide" or the phrase "weed killer," this search would only return documents that contain the word "safe" or the phrase "low risk" but also do not contain the words "danger" or "caution."


Executing the search

After entering the search terms, simply click the Search button near the top of the screen. The system will process your search, and if matches are found, you will be shown the Results List, described below.

If no matches to your search are found, you will be returned directly to the search entry screen with a message indicating the failure, and with your previous search string automatically entered into the text entry box. Adjust your search parameters to be more inclusive by adding terms or removing restrictions, and resubmit your search.

You may also be returned to the search term entry screen if there is a problem with the search processing. If this is the case, an error message will be displayed indicating what the problem was, and what action to take to fix it.


The Results List

Documents that match your search string are initially displayed in the results list. The list shows the number of matching documents, the rank of the matching document (Concept mode only), the document's number in the list, the document title, a document preview link, the size of the document in kilobytes, and the document type.

Additionally, selectors for choosing the displayed page of the results list (if more than one page is needed for the list), a button for performing a new search, and a button for searching through only the current matching documents are available.

Each of these features is described below:

Search Buttons

If the documents in the results list do not provide the information you need, or if you would like to perform a different search, simply click the New Search button at the top of the screen. You will return to the search entry screen and may start your search process over.

If the documents in the results list are relevant to your search, but there are too many of them, you can choose to refine your search by searching again only through the documents in this list. To do so, click the Search again through these results button. This will return you to the search entry screen, but the search you perform can be limited to only search through the previously-matched set of documents. You can continue to refine your search in this way an unlimited number of times. See the Refining Results section below for more information about refining your search.

Changing the Result Sorting

Each page of the results list displays the list in a tabular format, with table headings indicating the data shown in each column of the table. The Rank (Concept mode only), Title, and Size headings can each be clicked -- doing so changes the sorting of the results list to sorting based on the heading that was clicked. For Title and Size sorting, you can choose to sort the documents in ascending or descending order, by re-clicking the chosen table heading.

Ranking (Conceptual Search only)

Because Concept searches expand search terms beyond the exact word entered, the search software provides a mechanism for ranking how well it thinks a document matches the search. Ranking is not provided for Boolean searches, because by definition each document in a Boolean result set exactly matches the search string.

By default the results list is sorted by rank for Concept searches. This should place the "best" documents at the top of the list. A brief description of the ranking mechanisms used may help to explain why some documents rank higher than others for your search. The following items are all weighed against each other by the system when the document result set is ranked:

Completeness: The greater the number of query words (either exactly or expanded synonyms), the higher the rank. For example, if a document contains only one of two search terms, it can never rank higher than 50%.

Contextual Evidence: The greater the number of related terms, the higher the rank. For example, in a search for "birds" a document with many references to "pheasant," "duck," and "quail" may rank higher than a document with only one match for the exact term "birds."

Semantic Distance: The more closely related the terms, the higher the rank. Simply put, synonyms or other strongly-related terms will increase a document's rank, while weakly-related terms will decrease its rank.

Proximity: The closer together the query words and related terms within the document, the greater the rank. Search terms that appear near each other in a document increase the rank when compared to documents that contain the same terms spread evenly throughout the document.

Hit Density: The greater the ratio of query words and related terms to the total number of words in the document, the greater the rank. For example, a small document with ten matching terms may rank higher than a large document with 100 matching terms.

Viewing the Matching Document

To view a matching document, click on its title in the results list. The document will be displayed in a new window so that the results list is still available. For documents that are not standard HTML web pages, you will need to have a viewer for the file type installed. The most common non-HTML documents in the InfoBase are Adobe Acrobat PDF files, which require the free Acrobat Reader software to view. Other file types that the InfoBase supports include Corel WordPerfect, Microsoft Word, Microsoft Excel, and many others -- you will need the appropriate viewing application from the appropriate vendor to be able to view the original documents in those formats.

Tip for Adobe PDF viewing: The InfoBase will automatically highlight the matching search terms in the PDF document if your Adobe PDF viewer supports Adobe's Highlighting format (most of the recent versions of their software support this). If this is the case, you will see two icons in the Adobe Reader menu bar that look like this: highlight buttons If you see those buttons, you can use them to jump between the pages that contain the matches to your search, and the search terms you entered will display "highlighted" -- generally in the reverse colors of the rest of the text. This can be a great time saver for finding what you're looking for in large documents.

Note that most documents in this system will be retrieved by your computer from the original site, not from NPIC. There may be occasions where the original is not available, even though this search system shows that the document matches your search. If that is the case, or if you don't have the appropriate viewing software for the type of document, you can still retrieve much of the contents of the document from our Document Preview page, described below.

Viewing a Summary of Hits

If you would like to see why a document matches your search, or would like to preview the contents of a document before viewing the original, there are two ways to get more information about the document without actually opening the original. First, you may click the "plus" icon (+) to the right of the document's title in the results list. This will open a small summary area just below title of the corresponding document in the results list, and will provide you with the number of hits the document contains and a preview of each line in the document that has a hit.

If you would like to see the matching hits in the full context of the original document, you may click on any hit displayed in the summary area to view a converted version of the document with each hit highlighted (the Document Preview page). Once at this page, you may click on a displayed hit to jump to the next hit in the document.

You may also get to the document preview page by clicking on the Show link in the Document Preview column of the results table. In the Document Preview page, you will find details about the matching document (the document's Title, a link to the original document (if available), and the number of matching terms -- hits -- that occur), as well as specific context about the hits that occur in the document. Once you have viewed a Document Preview, a small green checkmark ()will appear next to the Show link in the results list to remind you that you've seen this document.

These Document Preview features may be used to view the contents of a document that is not currently available in original form. Please note that the version of the document displayed via this mechanism is as close as possible to the original, but it should not be considered a replacement or duplicate -- most notably any graphics or other non-textual information in the original is completely absent from this version. If looking for the authoritative source of this information, you must use the original document and not this version.


Refining results

From the results list, click the Search again through these results button to refine your existing search. Clicking this button will return you to the search entry screen, but the search you perform can be limited to only search through the previously-matched set of documents.

After clicking the Search again through these results button on the results list page, you are returned to the search entry screen, but it looks slightly different than the normal search screen. This version of the entry screen informs you of how many documents your previous search found -- you are now searching through this limited set of documents. There are also two different search buttons available, the new default Refine Existing Search button, and the optional New Search button. The former will run the search against only the previously-matched results. The latter allows you to run a "normal" new search through all chosen libraries.

You can continue to refine your search in this way an unlimited number of times. When refining searches, be careful to click the New Search button to return to "normal" mode when you are finished with your refined search. Entering an entirely new search string and running it against a previous result set will most likely not yield the expected results.