Blame - docs/0.8.8-incubating/aql/similarity.html - incubator-asterixdb-site

<h2><a name="Motivation_Back_to_TOC"></a><a name="Motivation" id="Motivation">Motivation</a> <font size="4"><a href="#toc">[Back to TOC]</a></font></h2>

224

<p>Similarity queries are widely used in applications where users need to find records that satisfy a similarity predicate, while exact matching is not sufficient. These queries are especially important for social and Web applications, where errors, abbreviations, and inconsistencies are common. As an example, we may want to find all the movies starring Schwarzenegger, while we don’t know the exact spelling of his last name (despite his popularity in both the movie industry and politics :-)). As another example, we want to find all the Facebook users who have similar friends. To meet this type of needs, AsterixDB supports similarity queries using efficient indexes and algorithms.</p></div>

225

226

<h2><a name="Data_Types_and_Similarity_Functions_Back_to_TOC"></a><a name="DataTypesAndSimilarityFunctions" id="DataTypesAndSimilarityFunctions">Data Types and Similarity Functions</a> <font size="4"><a href="#toc">[Back to TOC]</a></font></h2>

227

<p>AsterixDB supports <a class="externalLink" href="http://en.wikipedia.org/wiki/Levenshtein_distance">edit distance</a> (on strings) and <a class="externalLink" href="http://en.wikipedia.org/wiki/Jaccard_index">Jaccard</a> (on sets). For instance, in our <a href="primer.html#ADM:_Modeling_Semistructed_Data_in_AsterixDB">TinySocial</a> example, the <tt>friend-ids</tt> of a Facebook user forms a set of friends, and we can define a similarity between the sets of friends of two users. We can also convert a string to a set of grams of a length “n” (called “n-grams”) and define the Jaccard similarity between the two gram sets of the two strings. Formally, the “n-grams” of a string are its substrings of length “n”. For instance, the 3-grams of the string <tt>schwarzenegger</tt> are <tt>sch</tt>, <tt>chw</tt>, <tt>hwa</tt>, …, <tt>ger</tt>.</p>

228

<p>AsterixDB provides <a href="functions.html#Tokenizing_Functions">tokenization functions</a> to convert strings to sets, and the <a href="functions.html#Similarity_Functions">similarity functions</a>.</p></div>

229

230

<h2><a name="Similarity_Selection_Queries_Back_to_TOC"></a><a name="SimilaritySelectionQueries" id="SimilaritySelectionQueries">Similarity Selection Queries</a> <font size="4"><a href="#toc">[Back to TOC]</a></font></h2>

231

<p>The following query asks for all the Facebook users whose name is similar to <tt>Suzanna Tilson</tt>, i.e., their edit distance is at most 2.</p>

<pre> use dataverse TinySocial;

236

237

for $user in dataset('FacebookUsers')

238

let $ed := edit-distance($user.name, "Suzanna Tilson")

where $ed <= 2

return $user

</pre></div></div>

<p>The following query asks for all the Facebook users whose set of friend ids is similar to <tt>[1,5,9,10]</tt>, i.e., their Jaccard similarity is at least 0.6.</p>

<pre> use dataverse TinySocial;

247

248

for $user in dataset('FacebookUsers')

249

let $sim := similarity-jaccard($user.friend-ids, [1,5,9,10])

where $sim >= 0.6f

return $user

</pre></div></div>

<p>AsterixDB allows a user to use a similarity operator <tt>~=</tt> to express a condition by defining the similarity function and threshold using “set” statements earlier. For instance, the above query can be equivalently written as:</p>

<pre> use dataverse TinySocial;

258

259

set simfunction "jaccard";

260

set simthreshold "0.6f";

261

262

for $user in dataset('FacebookUsers')

263

where $user.friend-ids ~= [1,5,9,10]

264

return $user

265

</pre></div></div>

266

<p>In this query, we first declare Jaccard as the similarity function using <tt>simfunction</tt> and then specify the threshold <tt>0.6f</tt> using <tt>simthreshold</tt>.</p></div>

267

268

<h2><a name="Similarity_Join_Queries_Back_to_TOC"></a><a name="SimilarityJoinQueries" id="SimilarityJoinQueries">Similarity Join Queries</a> <font size="4"><a href="#toc">[Back to TOC]</a></font></h2>

269

<p>AsterixDB supports fuzzy joins between two sets. The following <a href="primer.html#Query_5_-_Fuzzy_Join">query</a> finds, for each Facebook user, all Twitter users with names similar to their name based on the edit distance.</p>

<pre> use dataverse TinySocial;

274

275

set simfunction "edit-distance";

276

set simthreshold "3";

277

278

for $fbu in dataset FacebookUsers

return {

"id": $fbu.id,

"name": $fbu.name,

"similar-users": for $t in dataset TweetMessages

283

let $tu := $t.user

284

where $tu.name ~= $fbu.name

285

return {

286

"twitter-screenname": $tu.screen-name,

287

"twitter-name": $tu.name

288

}

289

};

290

</pre></div></div></div>

291

292

<h2><a name="Using_Indexes_to_Support_Similarity_Queries_Back_to_TOC"></a><a name="UsingIndexesToSupportSimilarityQueries" id="UsingIndexesToSupportSimilarityQueries">Using Indexes to Support Similarity Queries</a> <font size="4"><a href="#toc">[Back to TOC]</a></font></h2>

293

<p>AsterixDB uses two types of indexes to support similarity queries, namely “ngram index” and “keyword index”.</p>

294

295

<h3><a name="NGram_Index"></a>NGram Index</h3>

296

<p>An “ngram index” is constructed on a set of strings. We generate n-grams for each string, and build an inverted list for each n-gram that includes the ids of the strings with this gram. A similarity query can be answered efficiently by accessing the inverted lists of the grams in the query and counting the number of occurrences of the string ids on these inverted lists. The similar idea can be used to answer queries with Jaccard similarity. A detailed description of these techniques is available at this <a class="externalLink" href="http://www.ics.uci.edu/~chenli/pub/icde2009-memreducer.pdf">paper</a>.</p>

297

<p>For instance, the following DDL statements create an ngram index on the <tt>FacebookUsers.name</tt> attribute using an inverted index of 3-grams.</p>

<pre> use dataverse TinySocial;

302

303

create index fbUserIdx on FacebookUsers(name) type ngram(3);

304

</pre></div></div>

305

<p>The number “3” in “ngram(3)” is the length “n” in the grams. This index can be used to optimize similarity queries on this attribute using <a href="functions.html#edit-distance">edit-distance</a>, <a href="functions.html#edit-distance-check">edit-distance-check</a>, <a href="functions.html#similarity-jaccard">similarity-jaccard</a>, or <a href="functions.html#similarity-jaccard-check">similarity-jaccard-check</a> queries on this attribute where the similarity is defined on sets of 3-grams. This index can also be used to optimize queries with the “<a href="functions.html#contains">contains()</a>” predicate (i.e., substring matching) since it can be also be solved by counting on the inverted lists of the grams in the query string.</p>

306

307

<h4><a name="NGram_Index_usage_case_-_edit-distance"></a>NGram Index usage case - <a href="functions.html#edit-distance">edit-distance</a></h4>

<pre> use dataverse TinySocial;

312

313

for $user in dataset('FacebookUsers')

314

let $ed := edit-distance($user.name, "Suzanna Tilson")

315

where $ed <= 2

316

return $user

317

</pre></div></div></div>

318

319

<h4><a name="NGram_Index_usage_case_-_edit-distance-check"></a>NGram Index usage case - <a href="functions.html#edit-distance-check">edit-distance-check</a></h4>

<pre> use dataverse TinySocial;

324

325

for $user in dataset('FacebookUsers')

326

let $ed := edit-distance-check($user.name, "Suzanna Tilson", 2)

327

where $ed[0]

328

return $ed[1]

329

</pre></div></div></div>

330

331

<h4><a name="NGram_Index_usage_case_-_similarity-jaccard"></a>NGram Index usage case - <a href="functions.html#similarity-jaccard">similarity-jaccard</a></h4>

<pre> use dataverse TinySocial;

336

337

for $user in dataset('FacebookUsers')

338

let $sim := similarity-jaccard($user.friend-ids, [1,5,9,10])

339

where $sim >= 0.6f

340

return $user

341

</pre></div></div></div>

342

343

<h4><a name="NGram_Index_usage_case_-_similarity-jaccard-check"></a>NGram Index usage case - <a href="functions.html#similarity-jaccard-check">similarity-jaccard-check</a></h4>

<pre> use dataverse TinySocial;

348

349

for $user in dataset('FacebookUsers')

350

let $sim := similarity-jaccard-check($user.friend-ids, [1,5,9,10], 0.6f)

351

where $sim[0]

352

return $user

353

</pre></div></div></div>

354

355

<h4><a name="NGram_Index_usage_case_-_contains"></a>NGram Index usage case - <a href="functions.html#contains">contains()</a></h4>

<pre> use dataverse TinySocial;

360

361

for $i in dataset('FacebookMessages')

362

where contains($i.message, "phone")

363

return {"mid": $i.message-id, "message": $i.message}

364

</pre></div></div></div></div>

365

366

<h3><a name="Keyword_Index"></a>Keyword Index</h3>

367

<p>A “keyword index” is constructed on a set of strings or sets (e.g., OrderedList, UnorderedList). Instead of generating grams as in an ngram index, we generate tokens (e.g., words) and for each token, construct an inverted list that includes the ids of the records with this token. The following two examples show how to create keyword index on two different types:</p>

368

369

<h4><a name="Keyword_Index_on_String_Type"></a>Keyword Index on String Type</h4>

<pre> use dataverse TinySocial;

374

375

drop index FacebookMessages.fbMessageIdx if exists;

376

create index fbMessageIdx on FacebookMessages(message) type keyword;

377

378

for $o in dataset('FacebookMessages')

379

let $jacc := similarity-jaccard-check(word-tokens($o.message), word-tokens("love like verizon"), 0.2f)

380

where $jacc[0]

381

return $o

382

</pre></div></div></div>

383

384

<h4><a name="Keyword_Index_on_UnorderedList_Type"></a>Keyword Index on UnorderedList Type</h4>

<pre> use dataverse TinySocial;

389

390

create index fbUserIdx_fids on FacebookUsers(friend-ids) type keyword;

391

392

for $c in dataset('FacebookUsers')

393

let $jacc := similarity-jaccard-check($c.friend-ids, {{3,10}}, 0.5f)

where $jacc[0]

return $c

</pre></div></div>

<p>As shown above, keyword index can be used to optimize queries with token-based similarity predicates, including <a href="functions.html#similarity-jaccard">similarity-jaccard</a> and <a href="functions.html#similarity-jaccard-check">similarity-jaccard-check</a>.</p></div></div></div>

</div>

</div>

</div>

<hr/>