HCDA/cikm-paper Changeset - 66e8d7b5ea2b · Centrum Wiskunde & Informatica (CWI)

@@ -198,11 +198,11 @@ and zoom in on \emph{only} the filtering step; and conduct an in-depth analysis

main components.  In particular, we study the effect of cleansing,

entity profiling, type of entity filtered for (Wikipedia or Twitter), and

document category (social, news, etc) on the filtering components'

performance. The main contribution of the

performance. The main contributions of the

paper are an in-depth analysis of the factors that affect entity-based

stream filtering, identifying optimal entity profiles without

compromising precision, describing and classifying relevant documents

that are not amenable to filtering , and estimating the upper-bound

that are not amenable to filtering, and estimating the upper-bound

of recall on entity-based filtering.

@@ -220,7 +220,7 @@ section \ref{sec:unfil}. Finally, we present our conclusions in

 \section{Data Description}\label{sec:desc}

We base this analysis on the TREC-KBA 2013 dataset%

Our study is carried out using the TREC-KBA 2013 dataset%

\footnote{\url{http://trec-kba.org/trec-kba-2013.shtml}}

that consists of three main parts: a time-stamped stream corpus, a set of

KB entities to be curated, and a set of relevance judgments. A CCR

@@ -234,14 +234,17 @@ is a  dump of  raw HTML pages. The cleansed version is the raw data

after its HTML tags are stripped off and only English documents

identified with Chromium Compact Language Detector

\footnote{\url{https://code.google.com/p/chromium-compact-language-detector/}}

are included.  The stream corpus is organized in hourly folders each

of which contains many  chunk files. Each chunk file contains between

hundreds and hundreds of thousands of serialized  thrift objects. One

thrift object is one document. A document could be a blog article, a

news article, or a social media post (including tweet).  The stream

corpus comes from three sources: TREC KBA 2012 (social, news and

linking) \footnote{\url{http://trec-kba.org/kba-stream-corpus-2012.shtml}},

arxiv\footnote{\url{http://arxiv.org/}}, and

are included.  The stream corpus is organized in hourly folders, each

of which contains many ``chunk files''. The chunk files contain

hunderds up to hundreds of thousands of semi-structured documents,

serialized as thrift objects (one thrift object corresponding to one

document). A document could be a blog article, a

news article, or a social media post (including tweet). The stream

corpus has been derived from three main sources: TREC KBA 2012

(containing blogs, news and web pages with shortened urls)%

\footnote{\url{http://trec-kba.org/kba-stream-corpus-2012.shtml}%

},

ArXiV\footnote{\url{http://arxiv.org/}}, and

spinn3r\footnote{\url{http://spinn3r.com/}}.

Table \ref{tab:streams} shows the sources, the number of hourly

directories, and the number of chunk files.

@@ -271,7 +274,11 @@ directories, and the number of chunk files.

\subsection{KB entities}

 The KB entities consist of 20 Twitter entities and 121 Wikipedia entities. The selected entities are, on purpose, sparse. The entities consist of 71 people, 1 organization, and 24 facilities.

The KB entities in the dataset consist of 20 Twitter entities and 121

Wikipedia entities. These selected entities are, on purpose, ``sparse

entities'' in the sense that they only occur in relatively few

documents. The collection of entities consists of 71 people entities,

1 organization entity, and 24 facilities entities.

\subsection{Relevance judgments}