Ju l 1 99 6 Morphological Analysis as Classification : an Inductive-Learning Approach | Semantic Scholar (2024)

It is concluded that lazy learning of morphological analysis as a classiication task is indeed a viable approach and has the strong advantages of avoiding the knowledge-acquisition bottleneck, being fast and deterministic in learning and processing, and being language-independent.

A method of using associative strength among morphemes, morpho-syntactic patterns, and syntactic categories to solve the ambiguities of segmentation and part-of-speech is proposed.

It is concluded that machine learning of morphological processing such as word form generation can be successfully learned in a supervised manner, without explicit description of rules.

A test bench for morphology learning is described, which would assist designers of morphology learning programs by providing both training and evaluation data, and would allow comparison across programs.

In this paper, the question why NLP is an interesting application for empirical ML is addressed, and a brief overview of current work in this area is provided.

The method proposed exploits corpora in order to extract words that are related morphologically between them and indicates their probability to have a morphological and semantic relation between them.

It is argued on the basis of computational modeling experiments that the adoption of abstractions (rules, patterns), taken as necessary for explaining generalization and productivity in both the generative and the cognitive grammar approach, is misguided.

The question why NLP is an interesting application for empirical ML is addressed, and a brief overview of current work in this area is provided.

It is found that in the early stage of training with a larger pool, more labeled examples are required to achieve a given level of accuracy than those with a smaller pool.

...

...

It is shown that using supervised learning techniques, based on a corpus of transcribed words, the same and even better performance can be achieved, without explicit modeling of linguistic knowledge.

The MOR-phological Parser MORPA is a fully implemented parser developed for use in a text-to-speech conversion system that combines a "conventional" context-free morphological grammar to filter out ungrammatical segmentations with a probability-based scoring function which determines the likelihood of each successful parse.

A data-oriented (empiricist) alternative to the currently pervasive (nativist) Principles and Parameters approach to the acquisition of stress assignment is investigated and a similarity-based algorithm is used to learn the system of main stress assignment in Dutch.

A particular performance-oriented approach to Natural Language Processing based on automatic memory-based learning of linguistic (lexical) tasks is proposed, and the consequences of the approach for computational lexicology are discussed.

The results indicate that the generalization performance of backpropagation learning for this task is not better than that of the best symbolic pattern matching approaches, and of exemplar-based generalization.

H hierarchical clustering techniques applied to NETtalk reveal that these different networks have similar internal representations of letter-to-sound correspondences within groups of processing units, which suggests that invariant internal representations may be found in assemblies of neurons intermediate in size between highly localized and completely distributed representations.

The architecture of the grapheme to phoneme conversion system is described and the solution to the alignment problem given the spelling and the phonetic trancription of a word is focused on.

A complete guide to the C4.5 system as implemented in C for the UNIX environment, which starts from simple core learning methods and shows how they can be elaborated and extended to deal with typical problems such as missing data and over hitting.

...

...