Analysing performance in a word prediction system with multiple prediction methods☆
Introduction
Like in many language technology applications in general (Dale et al., 2000, vii), the problem of word prediction boils down to finding a linguistically relevant mapping between the user input and system output. As user input, one or more word-initial characters, for instance, can be used in word completion, possibly matching a certain number of word tokens in the lexicon employed in the prediction system.
In practice, multiple prediction methods will have to be used in the same system to result in accurate enough predictions. Along with single word prediction, phrases consisting of two or more word tokens could be predicted in a hybrid prediction system with a phrase prediction utility. That kind of approach to word prediction is justified, because the token frequency of phrases appears to be sufficiently high in English texts in addition to being capable of modelling text structure as an alternation between single words and phrases (Erman and Warren, 2000). The German FASTY system developed by Matiasek et al. (2002) is an example of a prediction system with multiple prediction methods.
Regarding the quantification of word prediction performance, global measures of effort savings such as average keystroke or character savings percentages are common in present word prediction. A more detailed analysis of the functioning of different prediction techniques is, however, required from the point of view of technology. Apart from showing how various parameters of the prediction process proper actually affect the operation of a prediction system as a whole, an elaborated analysis of the performance of a novel prediction technique may reveal how that prediction method may be improved for its most optimal use.
A detailed analysis of the performance of a hybrid word prediction system with a phrase prediction utility alongside single word prediction is presented in the empirical part of this study, with a corpus of 383 lexical bundles and one text containing phrases of various kinds.
The following three parameters affect the performance of a hybrid prediction system with a phrase prediction utility: (1) cost of selecting an appropriate prediction mode for single word prediction and phrase prediction when it cannot be selected automatically based on the search key typed in; (2) phrase frequency in the predicted text, which must be sufficiently high to result in a significant performance improvement, say, about 5–10%; (3) coverage of the phrasal lexicon which must contain at least most of the phrases predicted. The hybrid prediction method presented here completes other prediction techniques such as a stipulated prediction method proposed by Shieber and Baker (2003), where a sequence of consonants are expanded into matching word tokens. Church and Thiesson (2005) suggest the use of wildcards in word prediction under degraded conditions.
The problem addressed in the experimental part of the study consists of determining the effect of the above three parameters, affecting the accuracy of a hybrid prediction system with a phrase prediction utility. To our knowledge, detailed analyses of parameters impacting on the prediction accuracy in that kind of prediction system are not yet reported in the literature, especially results regarding the cost of selecting an appropriate prediction technique.
The structure of the remaining part of the article is as follows: we first introduce word prediction in general terms, including basic prediction techniques currently used in commercial word prediction. Then we discuss measurement of word prediction performance from the point of view of the application user and technology, respectively. Section 4.2 attempts to show how the three parameters affect prediction accuracy and the operation of a hybrid prediction system as whole.
At the end of the article, we draw conclusions on the practical utility of phrase prediction as an additional prediction technique used together with single word prediction in a hybrid prediction system.
Section snippets
Word prediction
Word prediction systems aim at enhancing the rate of text production. By attempting to predict the single word or phrase the user is writing or intends to write next, the word prediction software aims to reduce the number of keystrokes or characters needed to type the text in. Monitoring the input letter-by-letter, a word prediction program associates a given input (search key) with choices of the word tokens available in the lexicon, producing a list of matching tokens. The contents of the
Corpora
Two test corpora were used to preliminarily evaluate the practical utility of phrase prediction in a hybrid prediction system. One test corpus consists of a collection of 383 phrases, representing lexical bundles taken from Biber et al. (1999), the other of one real text containing various kinds of phrases.
Lexical bundles are sequences of word forms (three or more tokens) which co-occur in natural discourse and appear regularly across a variety of texts, justifying their use as a test corpus
Results
In this section, we present preliminary results of the performance of a hybrid prediction system which attempts to model text structure in terms of alternation of single words and phrases within text. We assume that the prediction of entire phrases, i.e. phrase prediction, can improve the proportion of characters saved by 5–10% relative to single word prediction under optimal conditions.
Discussion
As might be expected, phrase frequency in the predicted text significantly affects the practical utility of the phrase prediction method. As seen in Fig. 3, phrase frequency will have to be 35% to result in an improved performance by about 6% points for the whole text, with 20% of overlapping keys and 98% coverage of the phrasal lexicon.
Given the maximum frequency of occurrence of phrases of 60% in written texts suggested by Erman and Warren (2000, p. 34), resulting in about 20% of phrase
Conclusions
In this article, we analysed in detail the performance of a hybrid prediction system with a phrase prediction utility along with single word prediction from the viewpoint of technology. An analysis of that kind is essential in a prediction system with multiple prediction methods for a number of reasons: first, it can explicitly show how different parameters of the prediction process proper actually affect the operation of the prediction system as a whole; second, it can also reveal how to
References (26)
- et al.
Longman Grammar of Spoken and Written English
(1999) - Boissière, P., Dours, D., 2002. An overview of existing writing assistance systems. In: Modelling, Measurement and...
- et al.
An overview of empirical Natural Language Processing
AI Magazine
(1997) - Carlberger, J., 1998. Design and implementation of a probabilistic word prediction program. Unpublished Master’s...
- et al.
Poisson mixtures
Journal of Natural Language Engineering
(1995) - Church, K., Thiesson, B., 2005. The Wild Thing! In: Proceedings of the ACL Interactive Poster and Demonstration...
- et al.
Assistive Technologies: Principles and Practice
(1995) - Copestake, A., 1997. Augmented and alternative NLP techniques for augmentative and alternative communication. Available...
Dictionary of Linguistic and Phonetics
(1997)
The idiom principle and the open choice principle
Text
Cited by (2)
Keystroke saving in a language with highly transparent orthography
2010, Journal of Computing and Information TechnologyAnalyzing statistical and syntactical English text for word prediction and text generation
2007, Information Technology Journal
- ☆
This article contains text and images that were previously published in a thesis available in Acta Universitatis Ouluensis Series. A full reference is given in references.