Update README.md

README.md

@@ -8,45 +8,39 @@ pinned: false
 ---
 
 
-# Random Baseline Model for Climate Disinformation Classification
+# Random Forest Model for Climate Disinformation Classification
 
 ## Model Description
 
-This is a random baseline model for the Frugal AI Challenge 2024, specifically for the text classification task of identifying climate disinformation. The model serves as a performance floor, randomly assigning labels to text inputs without any learning.
+This is a random forest model for the Frugal AI Challenge 2024, specifically for the audio classification task of identifying illegal deforestation.
 
 ### Intended Use
 
-- **Primary intended uses**: Baseline comparison for climate disinformation classification models
+- **Primary intended uses**: Illegal deforestation classification model
 - **Primary intended users**: Researchers and developers participating in the Frugal AI Challenge
 - **Out-of-scope use cases**: Not intended for production use or real-world classification tasks
 
 ## Training Data
 
-The model uses the QuotaClimat/frugalaichallenge-text-train dataset:
-- Size: ~6000 examples
+The model uses the rfcx/frugalai dataset:
+- Size: ~50000 examples
 - Split: 80% train, 20% test
-- 8 categories of climate disinformation claims
+- 2 categories of audio category
 
 ### Labels
-0. No relevant claim detected
-1. Global warming is not happening
-2. Not caused by humans
-3. Not bad or beneficial
-4. Solutions harmful/unnecessary
-5. Science is unreliable
-6. Proponents are biased
-7. Fossil fuels are needed
+0. chainsaw  (positively identifying a chainsaw)
+1. environment (not containing a chainsaw).
 
 ## Performance
 
 ### Metrics
-- **Accuracy**: ~12.5% (random chance with 8 classes)
+- **Accuracy**: ~89.3%
 - **Environmental Impact**:
   - Emissions tracked in gCO2eq
   - Energy consumption tracked in Wh
 
 ### Model Architecture
-The model implements a random choice between the 8 possible labels, serving as the simplest possible baseline.
+The model implements a random forest model used on pre-processed data. The pre-processing consists in a resampling, a Fourier decomposition and a standard scaler. 
 
 ## Environmental Impact
 
@@ -57,15 +51,9 @@ Environmental impact is tracked using CodeCarbon, measuring:
 This tracking helps establish a baseline for the environmental impact of model deployment and inference.
 
 ## Limitations
-- Makes completely random predictions
-- No learning or pattern recognition
-- No consideration of input text
-- Serves only as a baseline reference
-- Not suitable for any real-world applications
+- Takes some time to do the pre-processing.
 
 ## Ethical Considerations
 
-- Dataset contains sensitive topics related to climate disinformation
-- Model makes random predictions and should not be used for actual classification
 - Environmental impact is tracked to promote awareness of AI's carbon footprint
 ```