Construcción de una lista de palabras positivas y negativas

A lexicon is simply a list of words tagged with sentiment. It is the dictionary your rule-based detector will look words up in. 📖

Begin with two lists: positive words and negative words. Each list holds terms that reliably signal one feeling on their own.

Fill the positive set with words readers say when happy: good, great, love, excellent, amazing. Keep them lowercase for easy matching.

The negative set collects complaint words: bad, awful, hate, terrible, broken. These are the red flags your detector hunts for.

A Python set checks membership instantly, so testing if a word is positive stays fast even when your lists grow large.

Great and GREAT should match the same entry, so store words lowercased and lowercase incoming text too. This keeps matching consistent.

Looking a word up is one line. The in operator tells you instantly whether a term lives in your positive list.

You do not have to start blank. Tools like NLTK ship curated opinion lexicons with thousands of scored words you can reuse.

A word should not sit in both lists, so keep your sets disjoint. Conflicts make scoring ambiguous and confuse your final label.

Sentiment is domain-specific. For a hotel, quiet is positive; for a party, it is not. Tune the lexicon to your specific use case.

Treat your lists as living data. When the detector misses a clear cue, just add that word, so the lexicon steadily improves over time.

Think about why we chose this data structure.

You built a lexicon: positive and negative word sets, lowercased and disjoint, for fast lookups. You can borrow NLTK lists and tune them per domain. Next we score full reviews. ✅