sa_definition.tex 23.7 KB
 Loïc Barrault committed Mar 08, 2021 1 %!TEX root = text_processing_L1.tex  Loïc Barrault committed Nov 06, 2019 2 3 4 5 6 7 8  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{} \vfill \centering  Loïc Barrault committed Nov 10, 2019 9 \Huge{\edinred{[Sentiment Analysis]\\Motivations, definitions and more}}  Loïc Barrault committed Nov 06, 2019 10 11 12 13 14 15 16 17 18 19  \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{The world is digital!} \centering  Loïc Barrault committed Mar 08, 2021 20 \includegraphics[width=0.75\textwidth]{digital_around_world2020}  Loïc Barrault committed Nov 06, 2019 21   Loïc Barrault committed Mar 08, 2021 22 \source{https://wearesocial.com/blog/2020/07/digital-use-around-the-world-in-july-2020}  Loïc Barrault committed Nov 06, 2019 23 24 25 26 27 28 29 30 31 32 33 34  \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{The world is digital!} %\begin{block}{People express their \alert{emotions}, \alert{sentiments} or \alert{opinions}} People express their \alert{emotions}, \alert{sentiments} or \alert{opinions} \begin{itemize} \item comments on products (Amazon, Rakuten) \item comments on movies (Rotten Tomatoes, IMDB, Youtube)  Loïc Barrault committed Mar 08, 2021 35 \item experience in restaurants (Yelp!, Trip Advisor)  Loïc Barrault committed Nov 06, 2019 36 37 38 39 \item community websites (Facebook, Twitter, Instagram, LinkedIn, Reddit, Flickr) \end{itemize} %\end{block}  Loïc Barrault committed Mar 08, 2021 40 41 42 43 44 45 \begin{itemize} \item Large quantity of information \ra\ cannot be processed by humans \item[\ra] Companies wants business intelligence! \end{itemize}  Loïc Barrault committed Nov 06, 2019 46 47 48 49 50 51 52 53 54 \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{General goal of SA} %\begin{block}{Extract \alert{emotions}, \alert{sentiments} or \alert{opinions}} Extract \alert{emotions}, \alert{sentiments} or \alert{opinions} \begin{itemize}  Loïc Barrault committed Mar 08, 2021 55 \item Expressed by humans in \textbf{texts}  Loïc Barrault committed Nov 06, 2019 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 \item Use that information for business or intelligence purposes \item[\ra] Answer questions like: \begin{itemize} \item Do people liked this movie? \item Do people agree with this law? \item Customers: should you buy this washing machine? \item Prediction: who will win next election? \end{itemize} \end{itemize} %\end{block} \begin{block}{Sentiment analysis = opinion mining} \begin{itemize} \item considered as equivalent in our case... \item ...although a sentiment does not necessarily express an opinion! \end{itemize} \end{block} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Importance of SA} %\begin{block}{\alert{To make decisions!}} \alert{To make decisions!} \begin{itemize}  Loïc Barrault committed Nov 15, 2019 83 \item Politics: can replace surveys, polls, etc.  Loïc Barrault committed Nov 06, 2019 84 \item Market: complete/verify/correct vendor's advice  Loïc Barrault committed Nov 15, 2019 85 \item Television: decide wether to continue or stop a TV show  Loïc Barrault committed Nov 06, 2019 86 87 88 89 90 91 92 93 94 95 96 97 98 \item at a larger scale! \item[\ra] people express themselves on the web on any subject \item[\ra] provides an estimate of the global opinion \end{itemize} %\end{block} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Example: social media monitoring} \begin{center}  Loïc Barrault committed Nov 15, 2019 99 \includegraphics[width=0.6\textwidth]{twitter-sentiment-analysis}  Loïc Barrault committed Nov 06, 2019 100 101 \end{center}  Loïc Barrault committed Nov 15, 2019 102 \begin{center}  Loïc Barrault committed Nov 06, 2019 103 104 \source{https://www.heidislojewski.com/blog/2016/3/11/m301ztb9orke84cuhwc37j1wwumn6n} %https://www.kaggle.com/crowdflower/twitter-airline-sentiment and  Loïc Barrault committed Nov 15, 2019 105 \end{center}  Loïc Barrault committed Nov 06, 2019 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120  \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{An example + concepts} \begin{block}{An example} \only<1>{\myhl{white}{\myhl{white}{@JetBlue}, \myhl{white}{\myhl{white}{I} normally \myhl{white}{\includegraphics[width=.3cm]{heart} you}}, but this \myhl{white}{Late Flight flight experience} \myhl{white}{was the worst.}} \myhl{white}{2 hours on runway, no wifi \& tv not working properly}} \only<2>{\myhl{white}{\myhl{white}{@JetBlue}, \myhl{white}{\myhl{white}{I} normally \myhl{green!30}{\includegraphics[width=.3cm]{heart} you}}, but this \myhl{white}{Late Flight flight experience} \myhl{white}{was the worst.}} \myhl{white}{2 hours on runway, no wifi \& tv not working properly}} \only<3>{\myhl{white}{\myhl{white}{@JetBlue}, \myhl{lightgray}{\myhl{lightgray}{I} normally \myhl{green!30}{\includegraphics[width=.3cm]{heart} you}}, but this \myhl{white}{Late Flight flight experience} \myhl{white}{was the worst.}} \myhl{white}{2 hours on runway, no wifi \& tv not working properly}} \only<4>{\myhl{white}{\myhl{white}{@JetBlue}, \myhl{lightgray}{\myhl{lightgray}{I} normally \myhl{green!30}{\includegraphics[width=.3cm]{heart} you}}, but this \myhl{white}{Late Flight flight experience} \myhl{red!40}{was the worst.}} \myhl{white}{2 hours on runway, no wifi \& tv not working properly}} \only<5>{\myhl{white}{\myhl{cyan!40}{@JetBlue}, \myhl{lightgray}{\myhl{lightgray}{I} normally \myhl{green!30}{\includegraphics[width=.3cm]{heart} you}}, but this \myhl{blue!20}{Late Flight flight experience} \myhl{red!40}{was the worst.}} \myhl{white}{2 hours on runway, no wifi \& tv not working properly}}  Loïc Barrault committed Nov 15, 2019 121 122 123 %\only<6>{\myhl{white}{\myhl{cyan!40}{@JetBlue}, \myhl{lightgray}{\myhl{lightgray}{I} normally \myhl{green!30}{\includegraphics[width=.3cm]{heart} you}}, but this \myhl{blue!20}{Late Flight flight experience} \myhl{red!40}{was the worst.}} \myhl{white}{2 hours on runway, no wifi \& tv not working properly}} \only<6>{\myhl{white}{\myhl{cyan!40}{@JetBlue}, \myhl{lightgray}{\myhl{lightgray}{I} normally \myhl{green!30}{\includegraphics[width=.3cm]{heart} you}}, but this \myhl{blue!20}{Late Flight flight experience} \myhl{red!40}{was the worst.}} \myhl{orange!30}{2 hours on runway, no wifi \& tv not working properly}} \only<7>{\myhl{white}{\myhl{cyan!40}{@JetBlue}, \myhl{lightgray}{\myhl{brown!90}{I} normally \myhl{green!30}{\includegraphics[width=.3cm]{heart} you}}, but this \myhl{blue!20}{Late Flight flight experience} \myhl{red!40}{was the worst.}} \myhl{orange!30}{2 hours on runway, no wifi \& tv not working properly}}  Loïc Barrault committed Nov 06, 2019 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 \end{block} \begin{itemize} \item<2-> "\includegraphics[width=.3cm]{heart} you" is a positive \textbf{opinion} \item<3-> "I normally \includegraphics[width=.3cm]{heart} you" is a less positive \textbf{opinion} \item<4-> "was the worst." is a negative \textbf{opinion} \item<5-> "@JetBlue" and "Late Flight flight experience": the \textbf{target of opinions} %\item<6-> is the \textbf{target of opinion} \item<6-> "2 hours on runway, no wifi \& tv not working properly" is the \textbf{objective} part \item<7-> "I" is the \textbf{holder of opinion} \end{itemize} %567765664459669504 negative 1 Late Flight 0.6632 Delta gczark 0 @JetBlue, I normally ❤️ you, but this Late Flightst flight experience was the worst. 2 hours on runway, no wifi & tv not working properly 2015-02-17 11:20:57 -0800 New York, NY Eastern Time (US & Canada) \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Target of opinions} \begin{itemize} \item \textbf{Target of opinions} is a \myemph{person}, \myemph{event}, \myemph{organisation}, or \myemph{topic} \item Represented as a hierarchy of \myemph{components} having a set of \myemph{attributes} \begin{itemize} \item[\ra] let's call them both \myemph{features} \end{itemize} \item An \textbf{opinion} is expressed on any \myemph{feature} \end{itemize} \begin{textblock*}{60mm}[0,0](80mm,20mm) \includegraphics[width=0.95\textwidth]{target_opinions} \end{textblock*} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Factual vs. subjective data} \begin{block}{Facts} \begin{itemize} \item \emph{\textbf{something that is known to have happened or to exist, for which proof or information exists}} \item can be expressed using \emph{keywords} \item less/not subject to interpretation \item might be difficult to identify though! \end{itemize} \end{block} \begin{block}{Opinions / subjective data} \begin{itemize} \item \emph{\textbf{a thought, belief, judgment about something or someone}} \item no proof required \item can be hard to express with keywords \end{itemize} \end{block} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Factual vs. subjective data: examples} \begin{block}{What do people think of } \begin{itemize} % negative and/or positive? \item @SouthwestAir just did, thank you % negative \item @united which is why my next flights to Miami will be on another airline. % negative in the corpus \item @JetBlue please provide me your direct email for me to explain. % negative \item @united worst service ever \includegraphics[width=.3cm]{devil} \item It costed 500 dollars. \item It costed 500 dollars!!! \end{itemize} \end{block} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Subjectivity analysis} \textbf{Subjectivity classification} is often the first step for sentiment analysis. Aims: decide whether a text is \textbf{objective} of \textbf{subjective} \begin{itemize} \item \textbf{objective}: It costed 500 dollars. \item \textbf{subjective}: this Late Flight flight experience was the worst. \item[] \end{itemize} However, it is not always as simple as that... \begin{itemize} \item \textbf{objective} sentences can express opinion \myemph{indirectly}: \begin{itemize} \item \textit{My phone broke in the second day.} \end{itemize} \item \textbf{subjective} sentences do not always express positive of negative opinions: \begin{itemize} \item \textit{I think they will refund us.} \end{itemize} \end{itemize} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Bing Liu's model for Sentiment Analysis} An \textbf{opinion} is a quintuple $(o_j, f_{jk}, so_{ijkl}, h_i, t_l)$, where: \begin{itemize} \item $o_j$ is the targeted \myemph{object} of opinion. Also called \myemph{entity}. \item $f_{jk}$ is a \myemph{feature} of the object $o_j$. Also called \myemph{aspect}. \item $so_{ijkl}$ is the \myemph{sentiment value}\tikzmark{a} \item $h_i$ is the \myemph{sentiment holder} \item $t_l$ is the \myemph{time} \end{itemize} $so_{ijkl}$ can take several forms: \begin{itemize} \item positive, negative, neutral \item 1 to 5 stars, as in movie reviews \item more granular ratings. \end{itemize} \begin{textblock*}{50mm}[0,0](120mm,10mm) \tikzmark{b} \includegraphics[width=0.2\textwidth]{icons_vertical} \end{textblock*} \begin{tikzpicture}[overlay,remember picture] \draw[very thick, -Stealth] ($({pic cs:a})+(1ex,0ex)$) to [bend right, sloped, ""] ($({pic cs:b})+(-1ex,15ex)$); \end{tikzpicture} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Example} \only<1>{ \begin{block}{On 25th Oct. 2018, John Doe wrote: } “I bought the new iPhone a few days ago. It was such a nice phone. The touch screen was really cool. The voice quality was clear too. Although the battery life is not long, that is ok for me. However, my mother was mad with me as I did not tell her before I bought the phone. She also thought the phone was too expensive, and wanted me to return it to the shop.” \end{block}} %cyan!40 %green!30 %brown!90 \only<2>{ \begin{block}{On \myhl{orange}{25th Oct. 2018}, John Doe wrote: } “\myhl{brown!90}{I} bought the new \myhl{cyan!40}{iPhone} a few days ago. It was \myhl{green!30}{such a nice} \myhl{blue!20}{phone}. The \myhl{blue!20}{touch screen} was \myhl{green!30}{really cool}. The \myhl{blue!20}{voice quality} was \myhl{green!30}{clear} too. Although the \myhl{blue!20}{battery life} is \myhl{red!40}{not long}, that is ok for me. However, my mother was mad with me as I did not tell her before I bought the phone. \myhl{brown!90}{She} also thought the \myhl{blue!20}{phone} was \myhl{red!40}{too expensive}, and wanted me to return it to the shop.” \end{block} \begin{itemize} \item \myhl{cyan!40}{$o_j$}: iPhone \item \myhl{blue!20}{$f_{jk}$}: phone, touch screen, voice quality, battery life, phone \item $so_{ijkl}$: \myhl{green!30}{positive}, \myhl{green!30}{positive}, \myhl{green!30}{positive}, \myhl{red!40}{negative}, \myhl{red!40}{negative} \item \myhl{brown!90}{opinion holder $h_i$} : I, I, I, I, mother \item time \myhl{orange}{$t_l$} : 25/10/2018 \end{itemize} } \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis} \begin{block}{\textbf{Opinion mining}} \begin{itemize} \item Discover all quintuples $(o_j, f_{jk}, so_{ijkl}, h_i, t_l)$ in a document \end{itemize} \end{block} \begin{itemize} \item \textbf{Structure the unstructured} \begin{itemize} \item Use in data visualisation tools \item \myemph{Quantitative} and \myemph{qualitative} analysis \item[] \item To answer questions like: \begin{itemize} \item Do people like the new iPhone? \item Among them, what are the features that they most dislike? \item Among people that don't like the iPhone, what feature is missing? \end{itemize} \end{itemize} \end{itemize} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: not just ONE problem} Discover all quintuples $(o_j, f_{jk}, so_{ijkl}, h_i, t_l)$ in a document \begin{itemize} \item $o_j$ -- the target \myemph{object} or \myemph{entity}: \begin{itemize} \item \textbf{Named Entity Recognition}. \item \textbf{Coreference resolution}. \end{itemize} \item $f_{jk}$ -- a \myemph{feature} or \myemph{aspect}: \textbf{Information Extraction} (more about that later...). \item $so_{ijkl}$ -- \myemph{sentiment value}: \textbf{Sentiment Identification} \item $h_i$ -- \myemph{sentiment holder}: \textbf{Information/Data Extraction} \item $t_l$ -- \myemph{time}: \textbf{Information/Data Extraction} \end{itemize} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: granularity: document level} \begin{itemize} \item \myemph{Document} level \begin{itemize} \item Assumption: document focuses on a \textbf{single object} from a \textbf{single opinion holder}. \item Goal: discover $(\_,\_,so,\_,\_)$ \item[\ra] ignore $o_j, f_{jk}, h_i, t_l$ \item[] \item \textbf{Reviews} usually satisfy the assumption \begin{itemize} \item Positive = 4 or 5 \FiveStarOpen, Neutral = 3 \FiveStarOpen, Negative = 1 or 2 \FiveStarOpen \end{itemize} \end{itemize} \end{itemize} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: granularity: sentence level} \begin{itemize} \item<+-> \myemph{Sentence} level SA is similar to document level \begin{itemize} \item Assumption: sentence contains a \textbf{single opinion} from a \textbf{single opinion holder} \item Only an \textbf{intermediate} step \item Consists of two steps: \begin{enumerate} \item \myemph{Subjectivity} classification \Ra\ \textbf{detect} if the sentence expresses an opinion \item \myemph{Sentiment} classification \Ra\ identifies the sentence \textbf{polarity} \end{enumerate} \end{itemize} \end{itemize} \begin{enumerate} \item<+-> \myemph{Subjectivity} classification \begin{itemize} \item Bootstrapping method \cite{Riloff2003} \item Sentence similarity \cite{Yu2003} \item Naive Bayes \end{itemize} \item<+-> \myemph{Sentiment} classification \begin{itemize} \item Rule-based \item Corpus -based \item[\ra] detailed later... stay tuned! \end{itemize} \end{enumerate} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: granularity: feature level 1/2} \begin{block}{Motivation} \begin{itemize} \item Document and sentences may contain \textbf{mixed} opinions \item \textbf{Targets} of opinions (\myemph{features}) are ignored \item[\ra] SA at this level provides a \textbf{general} opinion on the \myemph{object} \item[\ra] Does not mean that \myemph{opinion holder} likes/dislikes everything about it. \end{itemize} \end{block} \begin{itemize} \item \myemph{Feature} level aims to provide a more fine-grained analysis \begin{itemize} \item \textbf{which} component people like/dislike \item[\ra] more informative analysis \item consists of 5 steps \end{itemize} \end{itemize} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: granularity: feature level 2/2} \begin{enumerate} \item<+-> Identify \myemph{entities} or \myemph{objects} \begin{itemize} %\item Similar to NER with some differences \item Given a set $Q$ of \textbf{seed entities} of class $C$, and a set $D$ of \textbf{candidate entities}, determine which of the entities in $D$ belong to $C$. \item[\ra] binary classification problem \end{itemize} \item<+-> Extract object \myemph{features} that have been commented on by the \myemph{opinion holder} \begin{itemize} \item Frequency-based methods \cite{Hu2004} \item Challenge: infrequent feature extraction \end{itemize} \item<+-> Group similar \myemph{features} \begin{itemize} \item \eg\ \emph{screen} and \emph{touch screen} \item \eg\ \emph{power usage} and \emph{battery life} \end{itemize} \item<+-> Classify the opinions as \green{\bf positive}, \orange{\bf neutral} or \edinred{\bf negative} \begin{itemize} \item[\Ra] detailed later \end{itemize} \item<+->{} [optional] Produce a summary of all feature-based opinions \end{enumerate} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: result} \begin{textblock*}{45mm}[0,0](100mm,-10mm) \begin{block}{{\fontsize{4}{2}\selectfont On \myhl{orange}{25th Oct. 2018}, John Doe wrote: }} \begin{spacing}{0.5} {\fontsize{4}{2}\selectfont “\myhl{brown!90}{I} bought the new \myhl{cyan!40}{iPhone} a few days ago. It was \myhl{green!30}{such a nice} \myhl{blue!20}{phone}. The \myhl{blue!20}{touch screen} was \myhl{green!30}{really cool}. The \myhl{blue!20}{voice quality} was \myhl{green!30}{clear} too. Although the \myhl{blue!20}{battery life} is \myhl{red!40}{not long}, that is ok for me. However, my mother was mad with me as I did not tell her before I bought the phone. \myhl{brown!90}{She} also thought the \myhl{blue!20}{phone} was \myhl{red!40}{too expensive}, and wanted me to return it to the shop.”} \end{spacing} \end{block} \end{textblock*} \begin{center} \begin{tabular}{p{2.5cm}cl} Component & Polarity & Comment \\ \toprule \textbf{Feature1} & \MR{3}{*}{\green{Positive} } & The touch screen was really cool. \\ \emph{screen} & & \small{The touch screen was so easy to use and can do amazing things.} \\ \emph{touch screen} & & The screen is so fluid! \\ \cmidrule{2-3} & \MR{2}{*}{\edinred{Negative} } & The screen is easily scratched. \\ & & \small{It's very difficult to remove finger marks from the touch screen.} \\ \midrule \textbf{Feature2} & \MR{2}{*}{\green{Positive}} & The battery stood for 2 days.\\ \emph{battery life} & & The battery lasts very long. \\ \cmidrule{2-3} & \MR{2}{*}{\edinred{Negative}} & I had to charge it after 2 hours... \\ & & ... \\ \bottomrule \end{tabular} \end{center} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: opinion observer} \only<1>{ \centering \includegraphics[width=0.8\textwidth]{sa_summary} Summary of reviews for smartphone 1 } \only<2>{ \centering \includegraphics[width=0.8\textwidth]{sa_summary_comp} Comparison of summaries of reviews for two smartphones } \cite{Liu2005} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: indicators} \begin{itemize} \item absolute number of expressed opinions and proportion \end{itemize} \centering \includegraphics[width=0.7\textwidth]{sa_opinion_proportion} \cite{Liu2011} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: indicators} \begin{itemize} \item evolution of opinions across time \end{itemize} \centering \includegraphics[width=0.7\textwidth]{sa_opinion_trend} \cite{Liu2011} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: challenges} \only<1>{ On \myhl{white}{4th Nov. 2018}, \myhl{white}{John Doe} wrote: \\ This past Saturday, \myhl{white}{I} bought a \myhl{white}{Nokia} phone and \myhl{white}{my girlfriend} bought a \myhl{white}{Motorola} phone with \myhl{white}{Bluetooth}. We called each other when we got home.\\ The \myhl{white}{voice} on \myhl{white}{my} phone \myhl{white}{was not so clear}, \myhl{white}{worse than my previous phone.}\\ The \myhl{white}{battery life} \myhl{white}{was short too.} \\ \myhl{white}{My girlfriend} \myhl{white}{was quite happy with her phone.}\\ \textbf{I wanted} a phone with good \myhl{white}{sound quality}. So \myhl{white}{my} \myhl{white}{purchase} \myhl{white}{was a real disappointment}. \\ I returned the phone yesterday. } \only<2->{ On \myhl{orange}{4th Nov. 2018}, \myhl{brown!90}{John Doe} wrote: \\ %\begin{spacing}{0.5}  Loïc Barrault committed Nov 29, 2019 562 This past Saturday, \myhl{brown!90}{\tikzmark{a0}I} bought a \myhl{cyan!40}{\tikzmark{a1}Nokia} phone and \myhl{brown!90}{\tikzmark{b0}my girlfriend} bought a \myhl{cyan!40}{\tikzmark{b1}Motorola} phone with \myhl{cyan!40}{Bluetooth}.  Loïc Barrault committed Nov 06, 2019 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 We called each other when we got home.\\ The \myhl{blue!20}{voice\tikzmark{a2}} on \myhl{brown!90}{my} phone \myhl{red!40}{was not so clear}, \myhl{red!40}{worse than my previous phone.}\\ The \myhl{blue!20}{battery life\tikzmark{a3}} \myhl{red!40}{was short too.} \\ \myhl{brown!90}{My girlfriend} \myhl{green!30}{was quite happy with her phone\tikzmark{b2}.}\\ \textbf{I wanted} a phone with good \myhl{blue!20}{sound quality}. So \myhl{brown!90}{my} \myhl{blue!20}{\tikzmark{a4}purchase} \myhl{red!40}{was a \tikzmark{a5}real disappointment}. \\ I returned the phone yesterday. \\ \vspace{1cm} \textbf{\ra\ Identify all the components of an opinion} } \only<3>{ \textbf{\ra\ Identify the relations} \begin{tikzpicture}[overlay,remember picture]  Loïc Barrault committed Nov 29, 2019 581  \draw[very thick, -Stealth, carminered] ($({pic cs:a0})+(1ex,2ex)$) to [bend left, sloped, ""] ($({pic cs:a1})+(1ex,+2ex)$);  Loïc Barrault committed Nov 06, 2019 582 583 584 585 586  \draw[very thick, -Stealth, carminered] ($({pic cs:a1})+(1ex,-1ex)$) to [bend left, sloped, ""] ($({pic cs:a2})+(1ex,+1ex)$); \draw[very thick, -Stealth, carminered] ($({pic cs:a1})+(1ex,-1ex)$) to [bend left, sloped, ""] ($({pic cs:a3})+(1ex,+1ex)$); \draw[very thick, -Stealth, carminered] ($({pic cs:a1})+(1ex,-1ex)$) to [bend left, sloped, ""] ($({pic cs:a4})+(3ex,+2ex)$); \draw[very thick, -Stealth, carminered] ($({pic cs:a4})+(3ex,2ex)$) to [bend left, sloped, ""] ($({pic cs:a5})+(2ex,+2ex)$);  Loïc Barrault committed Nov 29, 2019 587  \draw[very thick, -Stealth, carnationpink] ($({pic cs:b0})+(2ex,2ex)$) to [bend left, sloped, ""] ($({pic cs:b1})+(1ex,+1em)$);  Loïc Barrault committed Nov 06, 2019 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634  \draw[very thick, -Stealth, carnationpink] ($({pic cs:b1})+(3em,-1ex)$) to [bend left, sloped, ""] ($({pic cs:b2})+(1ex,+1ex)$); \end{tikzpicture} } \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{frame} \frametitle{Sentiment Analysis: Challenges} Discover all quintuples $(o_j, f_{jk}, so_{ijkl}, h_i, t_l)$ in a document \begin{itemize} \item $o_j$ -- the target \myemph{object} or \myemph{entity}: \begin{itemize} \item \textbf{Named Entity Recognition}. \item \textbf{Coreference resolution}. \end{itemize} \item $f_{jk}$ -- a \myemph{feature} or \myemph{aspect}: \textbf{Information Extraction} (more about that later...). \item $so_{ijkl}$ -- \myemph{sentiment value}: \textbf{Sentiment Identification} \item $h_i$ -- \myemph{sentiment holder}: \textbf{Information/Meta-Data Extraction} \item $t_l$ -- \myemph{time}: \textbf{Information/Meta-Data Extraction} \end{itemize} In addition: \begin{itemize} \item \textbf{Relation Extraction} \item \textbf{Synonym match}: e.g. "voice" == "sound quality" \end{itemize} \end{frame} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %\begin{frame} % \frametitle{Sentiment Analysis: principle} %\begin{block}{} % \begin{itemize} % \item % \end{itemize} %\end{block} % %\centering %\includegraphics[width=0.75\textwidth]{tp_concept} % %\end{frame}