Across the phases of validation, respondents reported 28 different first languages and interviews were conducted in 8 different languages (49.6% IsiZulu, 15.3% English, 12.8% isiXhosa, 6.4% Luganda, 6.3% SeSotho, 5.4% Runyoro, 3.9% Runyankole and 0.4% SeTswana) (N = 720; language data not collected for remaining 399 carers).
In subsequent analyses the magnitude and profile of this relationship remained similar for each language group (data not shown).
Social-media-language data, moreover, doesn't have to be used to study large groups; it can be usefully applied to individuals rather than communities.
Some of the questions that many people kept asking were 'how many pages are on the web?' and 'how many pages are there in each language?' Those data do not exist because there is no centralised unit - a web central intelligence unit of sorts.
This approach can be employed effectively to build language-independent STD systems, since prior knowledge of the language involved in the speech data is not necessary.
Other advantage is that QbE STD can be employed for building language-independent STD systems [7 10], since prior knowledge of the language involved in the speech data is not necessary.
On the other hand, QbE STD can be also employed for building language-independent STD systems [7, 8], which is mandatory when no or very limited training data are available to build a reliable speech recognition system, since a priori knowledge of the language involved in the speech data is not necessary.
This was a surprise for us because when we were porting Slavic languages, it was advantageous to use more data (even from relative languages and not the source or target one) as it guaranteed the robustness of the model.
According to the national guidelines, all consent documents in a research protocol need to be translated into the national language, irrespective of the language of data collection which may not be Amharic.
