Sentiment analysis of tweets in Polish language using deep...
summary
Detects five emotions — happiness anger sadness fear disgust — for Polish tweets using a pipeline that goes from collection to model training. Data come from Twitter via Twitter API and a scraper called getoftweetsfree, filtered by language with Lonnie Tech and lemmatized with Marcos. A 2,000‑word emotion lexicon maps words to 1–5 scores and sentence labels are means of word vectors; word embeddings use Polish Science Academy pretrained vectors. A model with embedding + LSTM(100) + sigmoid multi‑label output trains on ~400,000 tweets, tests on ~100, and reports ~82% train / ~79% test.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:10Agenda and research focus
- 2:19Collecting tweets in Polish
- 3:48Alternative method with GetOldTweets
- 4:32Data preprocessing
- 6:30Text representation challenges
- 7:37Pre-trained Polish language models
- 9:34Padding text sequences
- 10:42Emotion lexicon for Polish
- 12:12Converting words to grammar forms
- 13:31Handling synonyms in lexicon
- 14:49Emotional representation of tweets
- 15:36Rule-based emotion labeling results
- 16:03Neural network training
- 18:12Model performance metrics
- 18:36Q&A
related talks
37mMultilingual sentiment analysis & emotion detection based on social media data
39mThe art and science of making computers understand sentiment and emotion
01h 23mBeyond Sentiment Emotion Mining with Python and machine learning
01h 05mSocial Text, Sentiment, and Tone Analysis
32mDecoding Emotions from Art to Memes
43mSentiment Analysis in Action: Building Your Real-time Pipeline
43mGetting Sentimental: exploration of iOS sentiment analysis
39mSentiment Analysis, its techniques and applications