Update README.md

README.md

@@ -62,18 +62,6 @@ df = dataset.to_pandas()
 print(df.head())
 ```
 
-## Data Processing & Cleaning Pipeline
-
-The raw data contained significant noise, web scraping artifacts, and duplicate entries. A rigorous multi-step data cleaning pipeline was applied using Python, `datasets`, and `pandas`:
-
-1. **Noise Filtering:** Dropped completely empty or extremely short rows (less than 10 characters) that provide no context.
-2. **Artifact Removal:** Filtered out scraping artifacts, such as template errors starting with punctuation marks (e.g., `, İtalya'nın...`) and blocks consisting mainly of unreadable ASCII or newline characters.
-3. **Alphanumeric Ratio Check:** Rows with less than a 30% alphanumeric character ratio were discarded to eliminate symbol dumps.
-4. **Whitespace Normalization (Token Optimization):** 
-   - Reduced excessive consecutive newlines (3 or more) to standard double newlines (`\n\n`) to preserve paragraph structure without wasting the model's context window (tokens).
-   - Normalized multiple consecutive spaces/tabs into a single space.
-5. **Deduplication:** Converted to a Pandas DataFrame to perform exact string matching. A total of **2,796 duplicate texts** (mostly repetitive agency news and copied paragraphs) were permanently deleted to prevent model overfitting.
-
 ## Dataset Structure
 
 The dataset contains a single feature: