sentiment   2654

« earlier    

Stanford CoreNLP
Natural language software from Stanford for providing various lexical, syntactic, and semantic annotations for text.
nlp  textbook  treebanks  tagger  sentiment 
6 days ago by jerid.francom
SocialSent: Domain-Specific Sentiment Lexicons
The word soft may evoke positive connotations of warmth and cuddliness in many contexts, but calling a hockey player soft would be an insult. If you were to say something was terrific in the 1800s, this would probably imply that it was terrifying and awe-inspiring; today, terrific basically just implies that something is (pretty) good.

A word's sentiment or connotation depends on the domain or context in which it is used. However, previous computational work in natural language processing largely ignores this issue, and focuses and building and deploying generic domain-general sentiment lexicons.

SocialSent is a collection of code and datasets for performing domain-specific sentiment analysis. The SocialSent code package contains the SentProp algorithm for inducing domain-specific sentiment lexicons from unlabeled text, as well as a number of baseline algorithms.

We have also released domain-specific historical sentiment lexicons for 150 years of English and community-specific sentiment lexicons for 250 "subreddit" communities from The historical lexicons reveal that more than 5% of sentiment-bearing words switched their polarity from 1850 to 2000, and the community-specific lexicons highlight how sentiment varies drastically between online communities.
sentiment  datasets  data  nlp  code  programming  software  vocabulary  english 
15 days ago by msszczep
practioners guide Convolutional Neural Networks (CNNs) - Google Search
This article offers an empirical exploration on the use of character-level convolutional networks (ConvNets) for text classification. We constructed several largescale datasets to show that character-level convolutional networks could achieve state-of-the-art or competitive results. Comparisons are offered against traditional models such as bag of words, n-grams and their TFIDF variants, and deep learning models such as word-based ConvNets and recurrent neural networks.
CNN  text  classification  sentiment  analysis 
17 days ago by foodbaby
Three, six or 36: how many basic plots are there in all stories ever written? | Books | The Guardian
“There is no reason why the simple shapes of stories can’t be fed into computers; they are beautiful shapes,” said Vonnegut.

A new academic study has done exactly this, and gives us yet another reason to wish the great man were still with us to share his thoughts on it (and perhaps resubmit that thesis). Researchers from the University of Vermont’s Computational Story Lab fed 1,737 stories from Project Gutenberg – all English-language texts, all fiction – through a program that analysed their language for its emotional content.

Putting – maybe – an end to a debate that has been ongoing for millennia, the researchers found there are “six core trajectories which form the building blocks of complex narratives”. These are: “rags to riches” (a story that follows a rise in happiness), “tragedy”, or “riches to rags” (one that follows a fall in happiness), “man in a hole” (fall–rise), “Icarus” (rise–fall), “Cinderella” (rise–fall–rise), and “Oedipus” (fall–rise–fall). The most successful – here defined as the most downloaded – types of story, they find, are Cinderella, Oedipus, two sequential man in a hole arcs, and Cinderella with a tragic ending.

Their analysis (and the “simple shapes of stories” as theorised by Vonnegut) is provided online, and it’s fascinating to pick through. I liked the rise-fall-rise shape of Gulliver’s Travels, where words such as “destroy”, “enemy” and “ignorance” drag down the happiness rating, and the plunging “Icarus” graph of Romeo and Juliet, plagued by words such as “tears”, “die”, “weep” and “poison”.
literature  sentiment  analysis  nlp  text  story  stories 
4 weeks ago by msszczep
Sentiment lexicon for Portuguese
r  packages  sentiment  portuguese  data  textbook 
5 weeks ago by jerid.francom
Receptiviti - Home is a technology company with deep roots in academia that is using AI, NLP, Machine Learning and proprietary Language-Psychology Science to reinvent the way organizations understand and engage their most important assets -- people.
learning  machine  NLP  linguistics  sentiment  text 
6 weeks ago by kybernetikos
Machine learning bias
Sentiment classifier is bigoted. That's not a surprise; the problem is when it's applied wrong.
ml  machinelearning  google  sentiment  bias  racism  ai 
6 weeks ago by nelson

« earlier    

related tags

1896  1964  2015  2020  380  451  ai  algorithm  analysis  analytics  ant  api  apis  apple  architecture  art  atomictime  bias  bizalignment  books  c#  category  chimpom  classification  cloud  cloudnative  clustering  cnn  code  collapse  commentary  computervision  computing  conda  consumertaste  container  containers  coreos  crows  culture  cyanotype  data-analysis  data-science  data  datascience  dataset  datasets  datasources  dataviz  deep-learning  deeplearning  demolition  digital  dj  doc2vec  education  efficiency  emoji  emotion  encounter  english  ensemble  entity  eugeneonegin  eugenics  events  explanation  facebook  facedetection  fiction  financial  frequency  fulfillment  futurepresent  gartner  gender  globalism  google  gps  happiness  harrypotter  heaven  howto  idioms  if2017  ifttt  inference  infovis  inhabiting  insurance  internet  janeausten  journalism  keras  keyword  language  learn  learning  lexicon  linguistics  literature  love  lstm  machine-learning  machine  machinelearning  market_research  marketing  markov  may17  meaning  measurement  memory  memoryless  military  mirror  ml  modeling  momemtum  mood  movies  multi-cloud  music  narrative  network-analysis  neural-net  neuralnetworks  news  newsletter  nlp  nltk  notebook  nsa  numbers  olympics  orchistration  packages  paper  papers  plans  podemo  politics  portuguese  postar  present  programming  proposition  python  pytorch  quant  r-bloggers  r-project  r  racism  reddit  reference  reproduction  research  reviews  rnn  rstats  sarcasm  satellit  science  sentiment-analysis  sentimental  sentimentanalysis  socialmedia  society  software  spam  spectacle  speeches  starred  statistics  stocktwits  stories  story  stream  structure  surveillance  surveys  symbol  tagger  teaching  technology  television  text  textblob  textbook  theater  tidytext  time  tooling  tools  topic  topicmodel  trading  transformation  treebanks  troll  tutorial  tutorials  tweets  twitter  upsidedown  user  venn  video  visualization  vocabulary  wbm  weather  web  word2vec  wordembeddings  wordpress  workflow 

Copy this bookmark: