In this paper,  we hold the subsequent steps of the pipeline fixed, zoom in on the filtering step and  conduct an in-depth analysis of the main components in it.  In particular, we study  cleansing, different entity profiling,  type of entities (Wikipedia or Twitter), and type of documents (social, news, etc).  The main contribution of the paper: 

 An in-depth analysis of the factors that affect entity-based stream filtering

 Identifying optimal entity profiles vis-avis not compromising precision

 Describing relevant documents that are not amenable to filtering and thereby estimating the upper-bound on entity-based filtering

 The rest of the paper is is organized as follows: 

We use TREC KBA-CCR-2013 dataset \footnote{http://http://trec-kba.org/trec-kba-2013.shtml} and problem setting. The dataset consists of a time-stamped  stream corpus, a set of KB entities, and a set of relevance judgments. 

\subsection{Stream corpus} The stream corpus comes in two versions: raw and cleaned. The raw  and cleansed versions are 6.45TB and 4.5TB respectively,  after xz-compression and GPG encryption. The raw data is a  dump of  raw HTML pages. The cleansed version is the raw data after its HTML tags are stripped off and non-English docuemnts removed. 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{http://trec-kba.org/kba-stream-corpus-2013.shtml}, arxiv\footnote{http://arxiv.org/}, and spinn3r\footnote{http://spinn3r.com/}. Table \ref{tab:streams}   shows the sources, the number of hourly directories, and the number of chunk files. 

\begin{table*}

\caption{retrieved documents to different sources }

\begin{center}

 \begin{tabular}{l*{4}{l}l}

 documents     &   chunk files    &    Sub-stream \\

\hline

126,952         &11,851         &arxiv \\

\label{tab:streams}

\end{table*}

\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.  

\subsection{Relevance judgments}

TREC-KBA provided relevance judgments for training and testing. Relevance judgments are given to a document-entity pairs. Documents with citation-worthy content to a given entity are annotated  as \emph{vital},  while documents with tangentially relevant content, or docuemnts that lack freshliness o  with content that can be useful for initial KB-dossier are annotated as \emph{relevant}. Documents with no relevant content are labeled \emph{neutral} and spam is labeled as \emph{garbage}.  The inter-annotator agreement on vital in 2012 was 70\% while in 2013 it is 76\%. This is due to the more refined definition of vital and the distinction made between vital and relevant. 

 \begin{enumerate}

  \item Does cleansing affect filtering and subsequent performance

  \item What is the most effective way of entity profile representation

  \item Is filtering different for Wikipedia and Twitter entities?

  \item Are some type of documents easily filterable and others not ? 

  \item Does a gain in recall at filtering step translate to a gain in F-measure at the end of the pipeline?

  \item What are the vital(relevant documents that are not retrievable by a system?

  \item Are there vital (relevant) documents that are not filterable by a reasonable system?

\end{enumerate}

\subsection{Literature Review}

There has been a great deal of interest  as of late on entity-based filtering and ranking. One manifestation of that is the introduction of TREC KBA in 2012. Following that, there have been a number of research works done on the topic \cite{frank2012building, ceccarelli2013learning, taneva2013gem, wang2013bit, balog2013multi}.  These works are based on KBA 2012 task and dataset  and they address the whole problem of entity filtering and ranking.  TREC KBA continued in 2013, but the task underwent some changes. The main change between  the 2012 and 2013 are in the number of entities, the type of entities, the corpus and the relevance rankings.

The number of entities increased from 29 to 141, and it included 20 Twitter entities. The TREC KBA 2012 corpus was 1.9TB after xz-compression and had  400M documents. By contrast, the KBA 2013 corpus was 6.45 after XZ-compression and GPG encryption. A version with all-non English documented removed  is 4.5 TB and consists of 1 Billion documents. The 2013 corpus subsumed the 2012 corpus and added others from spinn3r, namely main-stream news, forum, arxiv, classified, reviews and meme-tracker.  A more important difference is, however, that the definition of the relevance ranking changed. The change   in the definitions of vital and relevant. While in KBA 2012, a document was judged vital if it has citation-worthy content, In 2013 it must have the freshliness, that is the content must trigger an editing of the KB entry. 

While the task of 2012 and 2013 are fundamentally the same, the approaches for the tasks varied due  to the size of the corpus. In the 2013, all participants used filtering to reduce the size of the big corpus.   They used different ways of filtering: many of them used two or more of different name variants from DBpedia such as labels, names, redirects, birth names, alias, nicknames, same-as and alternative names \cite{wang2013bit, dietzumass ,liu2013related, zhangpris}.  Although all of the participants used DBpedia name variants none of them used all them.  A few other participants used bold words in the first paragraph of the Wikipedia entity's profiles and anchor texts from other Wikipedia pages  \cite{bouvierfiltering, niauniversity}.  Very few participants used Boolean And built from the tokens of the canonical names \cite{illiotrec2013}.  

All of the studies mentioned used filtering as their first step to generate a smaller set of documents. And many systems suffered from poor recall and their system performances were highly affected \cite{frank2012building}. Although  systems  used different entity profiles to filter the stream, and achieved different performance levels, there is no study on and the factors and choices that affect the filtering step itself. Of course filtering has been extensively examined in TREC Filtering \cite{robertson2002trec}. However, those studies were isolated in the sense that they were intended to optimize recall. What we have here is a different scenario. Documents have relevance rating. Thus we want to study filtering in connection to  relevance to the entities and thus can be done by coupling filtering to the later stages of the pipeline. This is new to the best of our knowledge and the TREC KBA problem setting and data-sets offer a good opportunity to examine this aspect of filtering. 

Moreover, there has not been a chance to study at this scale and/or a study into what type of documents defy filtering and why? In this paper, we conduct a manual examination of the documents that are missing and classify them into different categories. We also estimate the general upper bound of recall using the different entities profiles and choose the best profile that results in an increased over all performance as measured by F-measure. 

\section{Method}

From Table \ref{tab:name} shows the difefrence between Wikipedia and Twitter entities.  Wikipedia entities' canonical names achieve a recall of 70\%, and partial names of canonical names achieve a recall of 86.1\%. This is an increase in recall of 16.1\%. By contrast, the increase in recall of partial names of all name variants over just all name variants is 8.3.  The high increase in recall when moving from canonical names  to their partial names, in comparison to the lower increase when moving from all name variants to their partial names can be explained by saturation. This is to mean that documents have already been extracted by the different name variants and thus using partial name does not bring in many new documents. One interesting observation is that, on Wikipedia entities, partial names of canonical names achieve better results than name variants.  This holds in both cleansed and raw extractions. %In the raw extraction, the difference is about 3.7. 

In Twitter entities, however, it is different. Both canonical and their partial names perform the same and the recall is very low. Canonical names and partial canonical names are the same for Twitter entities because they are one word names. For example in https://twitter.com/roryscovel, ``roryscovel`` is the canonical name and its partial name is also the same.  That they perform very low is because the canonical names of Twitter entities are not really names; they are usually arbitrarily created user names. It shows that  documents do not refer to Twitter entities by their user names. They refer to them by their display names, which is reflected in the recall (67.9\%). The use of partial names of all name variants increases the recall to 88.2\%.

When we talk at an aggregate-level (both Twitter and Wikipedia entities), we observe two important patterns. 1) we see that recall increases as we move from canonical names to canonical partial names, to all name variants, and to partial names of name variants. But we saw that that is not the case in Wikipedia entities.  The influence, therefore, comes from Twitter entities. 2) Using canonical names retrieves the least number of vital or relevant documents, and the partial names of all name variants retrieves the most number of documents. The difference in performance is 31.9\% on all entities, 20.7\% on Wikipedia entities, and 79.5\% on Twitter entities. This is a significant performance difference. 

The tables in \ref{tab:name} and \ref{tab:source-delta} show, recall for Wikipedia entities are higher than for Twitter. This indicates that Wikipedia entities are easier to match in documents than Twitter. This can be due to two reasons: 1) Wikipedia entities are relatively well described than Twitter entities. The fact that we can retrieve different name variants from DBpedia is a measure of relative description. By contrast, we have only two names for Twitter entities:their user names and their display names which we collect from their Twitter pages. 2) DBpedia entities are less obscure and that is why they are not in Wikipedia anyways. Another point is that mentioned by their display names more than they are by their user names. We also observed that social documents mention Twitter entities by their user names more than news suggesting a distinction between the standard in news and social documents. 

   \subsection{Impact on classification}

\begin{table*}

\begin{center}

\begin{tabular}{ll@{\quad}lllllll}

\hline

   all-entities &F& 0.241&0.261&0.259&0.265\\

	      &SU&0.259  &0.258 &0.263 &0.262 \\	

   Wikipedia &F&0.252&0.274& 0.265&0.271\\

	      &SU& 0.261& 0.259&  0.265&0.264 \\

   twitter &F&0.105&0.105&0.218&0.228\\

     &SU &0.105&0.250& 0.254&80.253\\

   &SU	     & 0.265&0.265 &0.266 & 0.259\\

   twitter&F &0.188&0.188&0.208&0.231\\

	&SU&    0.269 &0.250\% &0.250&0.253\\

\hline

\end{tabular}

\end{center}

\label{tab:class-vital}

\end{table*}

  \begin{table*}

\begin{center}

\begin{tabular}{ll@{\quad}lllllll}

\hline

   all-entities &F& 0.497&0.560&0.579&0.607\\

	      &SU&0.468  &0.484 &0.483 &0.492 \\	

   Wikipedia &F&0.546&0.618&0.599&0.617\\

   &SU&0.494  &0.513 &0.498 &0.508 \\

   twitter &F&0.142&0.142& 0.458&0.542\\

    &SU &0.317&0.328&0.392&0.392\\

\hline

   &SU	     & 0.483&0.498 &0.487 & 0.495\\

   twitter &F&0.210&0.210&0.499&0.580\\

	&SU&    0.319  &0.317 &0.421&0.446\\

\hline

\end{tabular}

\end{center}

\label{tab:class-vital-relevant}

\end{table*}

\subsection{Missing relevant documents \label{miss}}

There is a trade-off between using a richer entity-profile and retrieval of irrelevant documents. The richer the profile, the more relevant documents it retrieves, but also the more irrelevant documents. To put into perspective, lets compare the number of documents that are retrieved with partial names of name variants and partial names of canonical names. Using the raw corpus, the partial names of canonical names extracts a total of 2547487 and achieves a recall of 72.2\%. By contrast, the partial names of name variants extracts a total of 4735318 documents and achieves a recall of 90.2\%. The total number of documents extracted increases by 85.9\% for a recall gain of 18\%. The rest of the documents, that is 67.9\%, are newly introduced irrelevant documents. There is an advantage in excluding irrelevant documents from filtering because they confuse the later stages of the pipeline. 

 The use of the partial names of name variants for filtering is an aggressive attempt to retrieve as many relevant documents as possible at the cost of retrieving irrelevant documents. However, we still miss about  2363(10\%) the vital-relevant documents.  Why are these documents missed? If they are not mentioned by partial names of name variants, what are they mentioned by? Table \ref{tab:miss} shows the documents that we miss with respect to cleansed and raw corpus.  The upper part shows the number of documents missing from cleansed and raw versions of the corpus. The lower part of the table shows the intersections and exclusions in each corpus.  

\begin{table}

\caption{The number of documents that are missing from raw and cleansed extractions. }

\begin{center}

\begin{tabular}{l@{\quad}llllll}

\hline

\end{tabular}

\end{center}

\label{tab:miss}

\end{table}

 It is normal to assume that  the set of document-entity pairs extracted from cleansed are a sub-set of those   that are extracted from the raw corpus. We find that that is not the case. There are 217  unique entity-document pairs that are retrieved from the cleansed corpus, but not from the raw. 57 of them are vital.    Similarly,  there are  3081 document-entity pairs that are missing  from cleansed, but are present in  raw. 1065 of them are vital.  Examining the content of the documents reveals that it is due to a missing part of text from a corresponding document.  All the documents that we miss from the raw corpus are social, particularly from the category social  (not from weblogs). These are document such as tweets and other posts from other social media. To meet the format of the raw data, some of them must have been converted later after collection and on the way lost a part or all of their content. It is similar for the documents that we miss from cleansed: a part or the  content  is lost in 

converting.  In both cases the mention of the entity happened to be on the part of the text that is cut out during conversion. 

 The interesting set  of relevance judgments are those that  we miss from both raw and cleansed extractions. These are 2146 unique document-entity pairs, 219 of them are with vital relevance judgments.   The total number of entities in the missed vital annotations is  28 Wikipedia and 7  Twitter, making a total of 35. Looking into document categories shows that the  great majority (86.7\%) of the documents are social. This suggests that social (tweets and blogs) can talk about the entities without mentioning  them by name. This is, of course, inline with intuition. 

%    

%    \begin{table*}

% \caption{Breakdown of missing documents by sources for cleansed, raw and cleansed-and-raw}

% \begin{center}\begin{tabular}{l*{9}r}

%   &others&news&social \\

% \hline

% 

% 			&missing from raw only &   0 &0   &217 \\

% 			&missing from cleansed only   &430   &1321     &1341 \\

% 

% \end{center}

% \label{tab:miss from both}

% \end{table*}

We also observed that although docuemnts have different document ids, several of them have the same content. In the vital annotation, there are only three (88 news, and 409 weblog). In the 35  vital document-entity pairs we examined, and 13 are news and  22 are social. 

\section{Analysis and Discussion}

We conducted experiments to study  the impacts on recall of 

we conducted experiments to study the impacts of cleansing, 

\section{Conclusions}

In this paper, we examined the filtering step of the CCR pipeline by holding the classification stage constant. In particular, we studied the cleansing step, different entity profiles, type of entities (Wikipedia or Twitter), categories of documents (news, social, or others) and the vital (relevant) documents. While doing so, we attempted to find answers to these research questions: 

\begin{enumerate}

  \item Does cleansing affect filtering and subsequent performance

  \item What is the most effective way of entity profile representation

  \item Is filtering different for Wikipedia and Twitter entities?

  \item Are some type of documents easily filterable and others not ?