Re: [Dbpedia-discussion] Test Bulgarian mappings

I just opend https://github.com/dbpedia/extraction-framework/issues/289 On Mon, Dec 8, 2014 at 8:02 PM, Jona Christopher Sahnwaldt wrote: > Thanks for the notification. > > It's probably an encoding problem: extractionSamples/ doesn't seem to > work if there are any non-ASCII characters in the na

Re: [Dbpedia-discussion] object property extractor should check    rdfs:range

> usually more data is better (at least in a statistics-based approach). IMHO low quality data is never better. Don't know what you mean by "statistics-based approach", but if you mean Machine Learning, bad data is disastrous to learn from. Not always. If bad data is statistically negligible it

Re: [Dbpedia-discussion] Test Bulgarian mappings

Thanks for the notification. It's probably an encoding problem: extractionSamples/ doesn't seem to work if there are any non-ASCII characters in the name of the template. For example, Mapping_de:Infobox_Fluss works, but Mapping_de:Infobox_Straße doesn't: http://mappings.dbpedia.org/server/mapping

Re: [Dbpedia-discussion] question on statistics

Hi Vladimir, I don't know the details on how the numbers on the mapping wiki (part 2) are calculated. The numbers of the first part are calculated based on types and properties dumps. Due to the complexity of the extraction process, the relations between numbers you list won't hold. Below are s

Re: [Dbpedia-discussion] how to use ConstantMapping

Hi Vladimir, Thanks for reporting this. The "BAD URI" was a bug that should be fixed with [1]. (See inline) On Fri, Dec 5, 2014 at 5:34 PM, Vladimir Alexiev < vladimir.alex...@ontotext.com> wrote: > Hi! > > Should I ask usage questions on specific mapping templates here, or in the > wiki? > I'd

[Dbpedia-discussion] Test Bulgarian mappings

Hi, Last week I corrected some of the Bulgarian mappings connected with " ? ??", because they were wrong. How I can test the mappings because it seems that the links on Test this mapping

Re: [Dbpedia-discussion] object property extractor should check rdfs:range

> usually more data is better (at least in a statistics-based approach). IMHO low quality data is never better. Don't know what you mean by "statistics-based approach", but if you mean Machine Learning, bad data is disastrous to learn from. > 3. If you only need data that is compatible with the