preview

Method Of A Bottom Up Method And The Domain Of Digital Games

Good Essays

There is a growing number of research in which NLP practices are utilized. Ryan et al. (2015) applied techniques from NLP by employing a bottom-up approach to game studies. The authors introduced a LSA model that establishes the first application of a bottom-up method to the domain of digital games. LSA is a statistical method where words are ascribed semantic descriptions based on their contextual distributions across a large collection of text (corpus) (Landauer & Dumais, 1997). It determines how semantically related texts are to one another. LSA method is built on the assumption that words with similar meanings will occur in similar contexts and that related texts will be composed of similar words. This creates a co-occurrence matrix where each row represents a distinct term and each column a distinct document. The cells of the matrix are populated with frequency counts where each cell shows a count of the number of times the term of the corresponding row occurred in the document of the corresponding column. With large amount of terms appearing in a typical corpus, there is a high tendency of having large dimensional vectors with tens of thousands of entries. LSA can help reduced the dimensionality of these vectors by a variation of factor analysis called singular-value decomposition (SVD) (Golub & Reinsch, 1970). SVD allows researchers to specify the desired number of dimensions. It is important and often difficult to decide an appropriate number of dimensions for SVD

Get Access