Computing Skill NER

Nucha_SkillNER_BERT is a Named Entity Recognition (NER) model specifically fine-tuned to recognize skill-related entities from text, focusing on identifying both hard and soft skills. This model is built on top of a BERT-based architecture, allowing it to leverage contextual understanding for accurate extraction of skill-related information. It is particularly useful for analyzing job descriptions, resumes, or any text where skills are explicitly mentioned.

The model supports the recognition of multiple skill categories, including technical skills (e.g., programming languages, software tools) and soft skills (e.g., communication, leadership). It is ideal for applications in recruitment, talent management, or skill-based data analysis.

How to Use

You can use the Nucha/Nucha_SkillNER_BERT model for Named Entity Recognition (NER) by loading it directly from Hugging Face's transformers library. Below is an example of how to use the model with the pipeline API for entity extraction.

Step-by-Step Example:

# Libraly
from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline

# Load the pre-trained model and tokenizer
model_name = "Nucha/Nucha_SkillNER_BERT"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForTokenClassification.from_pretrained(model_name)

# Create a NER pipeline
ner_pipeline = pipeline("ner", model=model, tokenizer=tokenizer, aggregation_strategy="simple")

# Sample text
text = "I have experience in Python, JavaScript, and cloud technologies like AWS and Azure."

# Run the pipeline on the text
ner_results = ner_pipeline(text)

# Display the results
for entity in ner_results:
    print(f"Entity: {entity['word']}, Label: {entity['entity_group']}, Score: {entity['score']:.4f}")

Output Explanation:

Entity: This is the word or phrase identified in the text that matches one of the model's recognized categories.
Label: The classification label assigned to the entity, such as SKILL or TECHNOLOGY .
Score: The confidence score of the model for the identified entity, represented as a floating-point number.

Demo

The Nucha/Nucha_SkillNER_BERT model is designed for Named Entity Recognition (NER) specifically targeting skill-related entities in text. This demo allows users to input any text and see how well the model identifies different skills. https://huggingface.co/spaces/Nucha/NuchaSkillNER

How to Use:

Input Text: Enter any text that contains information about skills or related topics. For example, you can input job descriptions, resumes, or any relevant text.
Analyze: Click the "Analyze" button to run the model on the provided text. The model will process the input and extract named entities, specifically skills.
Results: The output will display the recognized entities along with their labels and confidence scores. The labels will indicate the type of skills identified (e.g., programming languages, frameworks, tools).

Evaluation

The Nucha/Nucha_SkillNER_BERT model has undergone rigorous evaluation to ensure its effectiveness in Named Entity Recognition (NER) tasks, specifically in identifying and categorizing skills relevant to various domains. The evaluation was conducted on a diverse set of datasets designed to reflect real-world scenarios.

Metrics

The model's performance was assessed using standard NER metrics:

Accuracy: Measures the overall correctness of the model's predictions.
Precision: Indicates the proportion of true positive results in the total predicted positives.
Recall: Reflects the ability of the model to find all relevant instances in the dataset.
F1 Score: The harmonic mean of precision and recall, providing a single score that balances both metrics.

                precision   recall    f1-score  support

      HSKILL    0.89        0.91      0.90      3708
      SSKILL    0.91        0.91      0.91      2299

   micro avg    0.90        0.91      0.90      6007
   macro avg    0.90        0.91      0.91      6007
weighted avg    0.90        0.91      0.90      6007

Accuracy: 0.9972517975663717      (Train:5083/Test:1017)

Testing Data

You can employ this model using the Transformers library's pipeline for NER, or incorporate it as a conventional Transformer in the HuggingFace ecosystem.

1017/5083

Results