Thesis (Ph.D. (Leer en Onderrig)--North-West University, Potchefstroom Campus, 2008.

An experimental comparison was made of two mediations of spelling instruction with sight words: a traditional paper and pencil presentation and a computer-assisted presentation. Five students identified as learning disabled participated in the investigation over the course of five to six weeks in an elementary mainstreamed setting. The microcomputer presentation consisted of a commercially available spelling program incorporating visual and auditory error verification procedures. A counterbalanced ABAC/ACAB intrasubject replication design was used to evaluate the spelling performances (percent correct, correct spelling sequences, and rates of responses).Findings1. The results indicated that the microcomputer presentation of the spelling words did not effectively enhance achievement over that of paper and pencil.2. The use of error verification procedures with the microcomputer format did not result in more efficient learning. Rather, these subjects learned more quickly without the verification procedures. Informal observations further indicated that the students ignored the cues provided for verification.3. The intrasubject replication format of this investigation indicated that there were no order effects of the treatment conditions or of the error verification conditions.4. The students in this experiment required a longer session each day and took more time to respond when working on the microcomputer than with the paper and pencil presentation. These students, however, lacked typing skills which increased their response times on the microcomputer.5. Informal observations indicated that attention-to-task behavior was enhanced by the microcomputer. Although the subjects worked for a longer period of time in this mode, their attention was focused on the task.6. Because of the questionable instructional value of much of the software, the use of microcomputers in comparison with traditional and less costly modes of instruction should be carefully evaluated.

In practice, most approaches for text error detection and correction are based on a conventional domain-dependent background dictionary that represents a fixed and static collection of correct words of a given language and, as a result, satisfactory correction can only be achieved if the dictionary covers most tokens of the underlying correct text. Again, most approaches for text correction are for only one or at best a very few types of errors. The purpose of this thesis is to propose an unsupervised approach to detecting and correcting text errors, that can compete with supervised approaches and answer the following questions: Can an unsupervised approach efficiently detect and correct a text containing multiple errors of both syntactic and semantic nature? What is the magnitude of error coverage, in terms of the number of errors that can be corrected? We conclude that (1) it is possible that an unsupervised approach can efficiently detect and correct a text containing multiple errors of both syntactic and semantic nature. Error types include: real-word spelling errors, typographical errors, lexical choice errors, unwanted words, missing words, prepositional errors, article errors, punctuation errors, and many of the grammatical errors (e.g., errors in agreement and verb formation). (2) The magnitude of error coverage, in terms of the number of errors that can be corrected, is almost double of the number of correct words of the text. Although this is not the upper limit, this is what is practically feasible. We use engineering approaches to answer the first question and theoretical approaches to answer and support the second question. We show that finding inherent properties of a correct text using a corpus in the form of an n-gram data set is more appropriate and practical than using other approaches to detecting and correcting errors. Instead of using rule-based approaches and dictionaries, we argue that a corpus can effectively be used to infer the properties of these types of errors, and to detect and correct these errors. We test the robustness of the proposed approach separately for some individual error types, and then for all types of errors. The approach is language-independent, it can be applied to other languages, as long as n-grams are available. The results of this thesis thus suggest that unsupervised approaches, which are often dismissed in favor of supervised ones in the context of many Natural Language Processing (NLP) related tasks, may present an interesting array of NLP-related problem solving strengths.

This study presents results from a project named The OVE-project (Orator Verbis Electris i.e. electric speech machine). The aim of this thesis is to describe and to evaluate a set of computer - programs based on synthetic speech. The programs are designed for training phonological awareness and are intended to be used as a remedial tool for children with reading and writing problems. During the evaluation and data-collecting period the project was economically and technically connected to the department of Speech, Music and Hearing at the Royal Institute of Technology in Stockholm. My associate, during this period, late Karoly Galyas has been in charge of the technical part of the project while I have been responsible for the content and the design of programs as well as for the evaluation and the assessment studies. The thesis has two overall aims. The first is to describe, analyze and discuss the development of a specific work process based on computer based supervising with synthetic speech feedback intended for children with reading and writing problems. The second overall aim is to describe, analyze and discuss the effect of the speech feedback and the specific work process on children with reading and writing problems. The evaluation and the data collection has been done within three different phases and consists of three training studies and two experimental studies: one text reproduction study and one spelling test study. The results showed very convincingly that the speech feedback had a positive effect on the pupils' self esteem. Activity and motivation were highly improved. A remarkable improvement of the pupils' ability to concentrate and persevere was observed. The results from the spelling test study showed high improvement rates when speech feedback in ongoing writing was used compared to handwriting and ordinary computer writing without speech feedback. The potentials of the synthetic speech feedback as a means to improve reading and writing skills is discussed. / digitalisering@umu

In practice, most approaches for text error detection and correction are based on a conventional domain-dependent background dictionary that represents a fixed and static collection of correct words of a given language and, as a result, satisfactory correction can only be achieved if the dictionary covers most tokens of the underlying correct text. Again, most approaches for text correction are for only one or at best a very few types of errors. The purpose of this thesis is to propose an unsupervised approach to detecting and correcting text errors, that can compete with supervised approaches and answer the following questions: Can an unsupervised approach efficiently detect and correct a text containing multiple errors of both syntactic and semantic nature? What is the magnitude of error coverage, in terms of the number of errors that can be corrected? We conclude that (1) it is possible that an unsupervised approach can efficiently detect and correct a text containing multiple errors of both syntactic and semantic nature. Error types include: real-word spelling errors, typographical errors, lexical choice errors, unwanted words, missing words, prepositional errors, article errors, punctuation errors, and many of the grammatical errors (e.g., errors in agreement and verb formation). (2) The magnitude of error coverage, in terms of the number of errors that can be corrected, is almost double of the number of correct words of the text. Although this is not the upper limit, this is what is practically feasible. We use engineering approaches to answer the first question and theoretical approaches to answer and support the second question. We show that finding inherent properties of a correct text using a corpus in the form of an n-gram data set is more appropriate and practical than using other approaches to detecting and correcting errors. Instead of using rule-based approaches and dictionaries, we argue that a corpus can effectively be used to infer the properties of these types of errors, and to detect and correct these errors. We test the robustness of the proposed approach separately for some individual error types, and then for all types of errors. The approach is language-independent, it can be applied to other languages, as long as n-grams are available. The results of this thesis thus suggest that unsupervised approaches, which are often dismissed in favor of supervised ones in the context of many Natural Language Processing (NLP) related tasks, may present an interesting array of NLP-related problem solving strengths.

Studien tar sin utgångspunkt i kognitiva skrivteorier. Syftet med studien har varit att undersöka om handskrivnings- stavnings- och avkodningsförmåga hade statistiska samband med textlängd och textkvalitetsamt om det fanns några skillnader mellan flickor och pojkar. Studien ville även se om det var bokstävernas läslighet eller om det var elevens föreställning om bokstavens utseende som hade betydelse för textlängd och textkvalitet. I urvalet ingick 38 elever i år 2. Resultaten visade att förmåga att snabbt skriva läsliga bokstäver och att kunna stava ord hade samband med textlängd. Textlängd hade samband med textkvalitet, ordvariation och syntaktisk komplexitet. Handskrivning och stavning hade ett måttligt samband med innehållsliga aspekter av textkvalitet och de förklarade 60 % av variationen på textkvalitet. Handskrivning och stavning hade starka samband med textyteaspekter av textkvalitet. Handskrivning och stavning hade dessutom ett inbördes samband. Avkodning hade samband med stavning och med textyteaspekter av textkvalitet. Flickor var bättre på handskrivning och de hade bättre läsbarhet i sina texter. Flickor hade också högre skrivhastighet, ett mer varierat ordval i texterna och använde stor bokstav och skiljetecken på ett säkrare sätt jämfört med pojkar. Inga könsskillnader fanns vad gällde stavning och innehållsliga aspekter av textkvalitet. Resultatet indikerar att handskrivning och stavning inte kan negligeras i undervisningen eftersom de påverkar textlängd och olika aspekter av textkvalitet.

Thesis (M.S.)--University of North Carolina Wilmington, 2009. / Title from PDF title page (February 17, 2010) Includes bibliographical references (p. 68-69)

Recent research in phonological awareness found a strong link between rhyming ability in preschool children and later reading achievement. The use of orthographic analogy, the ability to make inferences from similarities in spelling to similarities in sound, was proposed as the mechanism to explain this relationship (Goswami & Bryant, 1990). Literature was presented that suggested the need for further research. Four research questions were examined. First, can prereaders learn to read unfamiliar words on the basis of orthographic analogy after brief training with rhyming words? The evidence supported the view that they could. Second, will the ability to read words by orthographic analogy be enhanced by phonological training in onset and rime, and by the use of segmented text? The brief phonological training did not increase analogy word reading over the same training without it. However, using text segmented at the onset-rime boundary for training items did increase analogy word reading. Third, will reading by orthographic analogy vary according to the level of prereading skills (rhyming ability, phoneme identity, letter-sound knowledge)? The majority of children with high prereading skills learned to read analogy test words whereas most children with low prereading skills found the task too arduous. Fourth, will rhyming ability make an independent contribution to reading achievement? The results were equivocal. Rhyming ability did make an independent contribution to the number of trials taken to learn the training items. It did not when analogy word reading was the dependent variable. Phoneme identity accounted for most of the variance in analogy word reading. Further analyses found that the ability to identify the final phoneme was the best discriminator between children who learned to read analogy test words and those who did not. A possible explanation was that children used the final phoneme to determine the sound of the rime ending rather than the last two phonemes together.

This work addresses issues in the automatic preprocessing of historical German input text for use by conventional natural language processing techniques. Conventional techniques cannot adequately account for historical input text due to conventional tools' reliance on a fixed application-specific lexicon keyed by contemporary orthographic surface form on the one hand, and the lack of consistent orthographic conventions in historical input text on the other. Historical spelling variation is treated here as an error-correction problem or "canonicalization" task: an attempt to automatically assign each (historical) input word a unique extant canonical cognate, thus allowing direct application-specific processing (tagging, parsing, etc.) of the returned canonical forms without need for any additional application-specific modifications. In the course of the work, various methods for automatic canonicalization are investigated and empirically evaluated, including conflation by phonetic identity, conflation by lemma instantiation heuristics, canonicalization by weighted finite-state rewrite cascade, and token-wise disambiguation by a dynamic Hidden Markov Model. / Diese Arbeit behandelt Themen der automatischen Vorverarbeitung historischen deutschen Textes für die Weiterverarbeitung durch konventionelle computerlinguistische Techniken. Konventionelle Techniken können historischen Text wegen des hohen Grads an graphematischer Variation in solchem Text ohne eine solche Vorverarbeitung nicht zufriedenstellend behandeln. Variation in der historischen Rechtschreibung wird hier als Fehlerkorrekturproblem oder "Kanonikalisierungsaufgabe" behandelt: ein Versuch, jedem (historischen) Eingabewort eine eindeutige extante Äquivalente zuzuordnen; so können konventionelle Techniken ohne weitere Modifikation direkt auf den gelieferten kanonischen Formen arbeiten. Verschiedene Methoden zur automatischen Kanonikalisierung werden im Rahmen dieser Arbeit untersucht, unter anderem Konflation durch phonetische Identität, Konflation durch Lemma-Instanziierungsheuristiken, Kanonikalisierung durch eine Kaskade gewichteter endlicher Transduktoren, und Disambiguiierung von Konflationskandidaten durch ein dynamisches Hidden Markov Modell.

We examined the relationship between orthographic processing and spelling in French immersion children. Study 1 included 148 first graders and they were assessed on orthographic processing and spelling in English and French. In Study 2, we followed 69 second graders for two years. Orthographic processing and spelling in English and French were administered in second and third grade. In Study 3, we analyzed the spelling errors made by the third graders in Study 2. In Study 1, we found a within-language relationship in English and French between orthographic processing and spelling. Cross-language transfer from French orthographic processing to English spelling was also observed. In Study 2, Grade 2 English spelling predicted gains in Grade 3 English and French orthographic processing. Study 3 showed that children made transfer errors when spelling in English and French. Overall, the current research highlights the importance of orthographic processing and spelling in French immersion children.