No description available.

No description available.

No description available.

This thesis evaluates the translation quality of phrase-based machine translation system. It explains the translation error annotation scheme to manually annotate errors related to English to Urdu translation system. The primary goal of the thesis is to experiment with different heuristic in order to improve the translation quality based on through manual analysis of 200 test sentences. Different hueristics such as (1) pre-processing of input English, such as word reordering, (2) preprocessing the training corpus in order to improve word alignment, (3) using additional factors (in Moses factored translation) to better model target-side morphological coherence are applied and their impact on the translation quality is evaluated.

Background: Grammars of Old English held at least until the 1960s that word orderin Anglo-Saxon texts was essentially “free”, that is, determined entirely or primarily by stylistic choice rather than syntactic rules.  Although prose word order has been shown to be regular in several models, the same cannot be said of poetry.  This study uses Nils-Lennart Johannesson’s Old English syntax model, operating within the Government and Binding framework, to establish whether the phrase structure of The Wanderer can fit into this model as it stands, and if not, whether a reasonably small number of additional parameters can be established in order to establish whether “free” word order is in evidence, or whether the word order of Old English poetry is regular in the same way as prose. Results: A full clause analysis showed that the majority of the clauses fit Johannesson’s model.  For those which did not, two modifications are recommended: non-compulsory movement of main verbs in main clauses from I to C; and the splitting and rightwards extraposition of the second part of coordinated NPs in which the first coordinated element is “light” and the second “heavy”.  This leaves a small number of clauses featuring constructions which do not occur frequently enough in the text to allow rules to be induced to explain them.  These must therefore be deemed irregular.  Conclusions:  While much of The Wanderer has been shown to be syntactically regular, some constructions could not be fitted into the existing model without the introduction of special parameters to excuse them.  This paper is intended as a pilot study for a larger project which will incorporate the other poems in the heroic tradition with the hope of inducing a complete syntax for them.  One part of that investigation will be to include these infrequent constructions in The Wanderer, to find comparable constructions in other poems and categorise them within the corpus.