Hal ini mengutip dari jenis pengukuran yag dilakukan oleh Simon J. Bell di tahun 2006 dari University of Melbourne dalam publikasinya yang berjudul Coping With Customer Complaints

How to use

You can use this model directly with a pipeline for masked language modeling:

python

>>>pip install torch
>>>pip install transformers
>>>from transformers import BertTokenizer, BertForSequenceClassification
>>>import torch

model_name = "radityapranata/absabert-keluhanpln-v3"
tokenizer = BertTokenizer.from_pretrained(model_name)
model = BertForSequenceClassification.from_pretrained(model_name)
absa_pipeline = pipeline("sentiment-analysis", model=model, tokenizer=tokenizer)

text = "informasi tambah daya" #"Token listrik tidak dapat terisi, kwh meter tulisan periksa."
inputs = tokenizer(text, return_tensors="pt")
result = absa_pipeline(text)

outputs = model(**inputs)
logits = outputs.logits

get_aspect = torch.argmax(logits, dim=1).item()
aspects = ["Layanan Pelanggan", "Ketersediaan Produk", "Kebijakan Usaha", "Pemulihan Layanan"]
aspect = aspects[get_aspect]

for aspect_result in result:
    Sentiment = aspect_result["label"]
    Score = aspect_result["score"]
    print("Aspect:", aspect)
    print(f"Sentiment: {Sentiment}, Score: {Score}")