Prashant Vishwakarma
Statistical modeling, evidence-based insights.
New to Contra
Prashant is ready for their next project!
30-day price forecasting using seasonality and trend decomposition. Backtested against real Google Finance data. Metrics: MAPE 4.2%, R² 0.89.
0
5
:Unsupervised anomaly detection on IoT sensor data using LSTM autoencoder reconstruction error with learned threshold. Metrics: Detection Rate 96.1%, F1 0.94, False Alarm Rate 2.3%.
0
9
Fine-tuned bert-base-uncased on 50K product reviews for 3-class sentiment classification (Positive / Neutral / Negative). Metrics: Accuracy 92.4%, Macro F1 0.91.
0
14
Built a Residual CNN for image super-resolution (PSNR 32.4 dB, SSIM 0.91) with Grad-CAM interpretability to visualize per-layer spatial feature reconstruction.
0
20