nlp-lab/Project/naive_approach
2018-07-27 13:15:30 +00:00
..
Emoji_Distance.ipynb user interface 2018-06-19 16:28:26 +02:00
Emoticon_Emoji_Assignment.ipynb emoticon_emoji_assignment_progeamm and resulting file 2018-06-05 14:33:15 +02:00
fastTextVectors.kv FastText eingebunden 2018-07-21 15:15:04 +02:00
naive_approach.ipynb updated naive approach 2018-06-05 15:03:33 +02:00
naive_approach.py stemming als parameter und anpassungen für evaluation 2018-07-23 12:09:36 +02:00
naiveApproachTest.ipynb naive approach test 2018-06-26 15:12:07 +02:00
README.md Update 'Project/naive_approach/README.md' 2018-07-27 13:15:30 +00:00
word2vec.model Naive approach mit word2Vec similarities 2018-07-07 18:25:17 +02:00

naive_approach

This directory contains the functions necessary to run the Naive Approach.

Prerequisites:

For testing, import naive_approach.py and execute the following commands:

  1. prepareData(stem, lower)

    • preprocesses the emoji descriptions and returns a dictionary with the indexed emojis
    • parameters:
      • stem: Apply stemming (default=True)
      • lower: Apply lowercasing (default=True)
  2. predict(sentence, lookup, emojis_to_consider, criteria, lang, embeddings, n=10, t=0.9)

    • evaluates an input sentence and returns a list of predicted emojis
    • parameters:
      • sentence: Input sentence (required parameter)
      • lookup: dictionary with emoji data (return value of prepareData, required parameter)
      • emojis_to_consider: set of emojis to include in prediction, or "all" (default="all")
      • criteria: criteria to evaluate the values of the description - message matching.
        • options: "sum", "mean", "max_val", "threshold" (default: "threshold")
      • lang: language to use (default: "eng")
      • embeddings: word embeddings
        • options: "wordnet", "word2Vec", "fastText", default: "wordnet"
      • n: number of top ranked emojis to return (default=10)
      • t: threshold for the "threshold" criteria (default=0.9)