Identifying inadvertent semantic errors in English texts.
Material type:
TextDescription: 91 pISBN: 0599815906Subject(s): Computer Science | Language, Linguistics | 0984 | 0290Dissertation note: Thesis (Ph.D.)--The University of Texas at Dallas, 2000. Summary: Define an inadvertent semantic error to be a spelling or typing error that turns an intended word into another word of the language. For example, the intended word “sight” might become the word “site.” A spell checker cannot identify such an error. In the English language—the case of interest here—a syntax checker may also fail to catch such an error since the parts-of-speech of an erroneous word may permit an acceptable parsing. In addition, error detection by a syntax checker likely is difficult if the text contains many special terms, symbols, formulas, or conventions whose syntactic contributions cannot be established without a complete understanding of the text. For such texts, as well as for texts that do not involve such complicating aspects, this dissertation presents an effective technique for identifying the majority of inadvertent semantic errors.Summary: The method has been added to an existing software system for spell and syntax checking. We have conducted tests involving mathematical book chapters, technical papers, and newspaper texts in two subject areas. To judge the accuracy of the method, we focused on the words of testing texts whose usage was reasonably well represented in the prior texts. For such cases, the method on average found 72% of inadvertent semantic errors in large texts and 87% of such errors in small texts.Summary: In the implemented system, each likely error is pointed out to the user, who then either makes an appropriate correction or declares that an error actually is not present. The latter case of false-bad diagnosis is annoying to the user and should be avoided. In the tests, that mistake did not occur often. Indeed, false-bad diagnoses occurred on average for 23 word instances of a large text and for I word instance of a small text.Summary: These statistics show that the method is remarkably effective for the recognition of inadvertent semantic errors.*Summary: *This research was supported in part by the Office of Naval Research under Grant N00014-93-1-0096.
Source: Dissertation Abstracts International, Volume: 61-06, Section: B, page: 3126.
Supervisor: Klaus Truemper.
Thesis (Ph.D.)--The University of Texas at Dallas, 2000.
Define an inadvertent semantic error to be a spelling or typing error that turns an intended word into another word of the language. For example, the intended word “sight” might become the word “site.” A spell checker cannot identify such an error. In the English language—the case of interest here—a syntax checker may also fail to catch such an error since the parts-of-speech of an erroneous word may permit an acceptable parsing. In addition, error detection by a syntax checker likely is difficult if the text contains many special terms, symbols, formulas, or conventions whose syntactic contributions cannot be established without a complete understanding of the text. For such texts, as well as for texts that do not involve such complicating aspects, this dissertation presents an effective technique for identifying the majority of inadvertent semantic errors.
The method has been added to an existing software system for spell and syntax checking. We have conducted tests involving mathematical book chapters, technical papers, and newspaper texts in two subject areas. To judge the accuracy of the method, we focused on the words of testing texts whose usage was reasonably well represented in the prior texts. For such cases, the method on average found 72% of inadvertent semantic errors in large texts and 87% of such errors in small texts.
In the implemented system, each likely error is pointed out to the user, who then either makes an appropriate correction or declares that an error actually is not present. The latter case of false-bad diagnosis is annoying to the user and should be avoided. In the tests, that mistake did not occur often. Indeed, false-bad diagnoses occurred on average for 23 word instances of a large text and for I word instance of a small text.
These statistics show that the method is remarkably effective for the recognition of inadvertent semantic errors.*
*This research was supported in part by the Office of Naval Research under Grant N00014-93-1-0096.
School code: 0382.
There are no comments on this title.