eligapris/lie-detection-sentiment-analysis

#This model is designed to identify and classify text into number of categories:

It leverages advanced Natural Language Processing (NLP) techniques, specifically sentiment analysis, to determine the overall attitude or opinion expressed within a piece of text. By combining this with a dedicated dataset focusing on identifying lies and fakes, it aims to accurately predict whether a given statement is true or false.

[
  [
    {
      "label": "half-true",
      "score": 0.21052952110767365
    },
    {
      "label": "mostly-true",
      "score": 0.19538265466690063
    },
    {
      "label": "false",
      "score": 0.1879868507385254
    },
    {
      "label": "barely-true",
      "score": 0.16795198619365692
    },
    {
      "label": "true",
      "score": 0.1583855301141739
    },
    {
      "label": "pants-fire",
      "score": 0.0797634944319725
    }
  ]
]

Model Trained Using AutoTrain

• Problem type: Text Classification

Validation Metrics

loss: 1.757171869277954

f1_macro: 0.05706191825171995

f1_micro: 0.20654296875

f1_weighted: 0.07071442798968029

precision_macro: 0.034423828125

precision_micro: 0.20654296875

precision_weighted: 0.04265999794006348

recall_macro: 0.16666666666666666

recall_micro: 0.20654296875

recall_weighted: 0.20654296875

accuracy: 0.20654296875