Articles

Cross-Country Transfer Learning for FDI Forecasting in Small Macroeconomic Datasets

/0 Comments/in /by

This study examines whether transfer learning can improve machine learning performance under data-scarce forecasting conditions. The empirical application focuses on forecasting foreign direct investment (FDI) inflows using CatBoostRegressor. The source domain is represented by a broad international panel of countries, while the target domain consists of Central Asian economies.

CatBoost is evaluated in two settings: without transfer learning, where the model is trained only on target-region data, and with transfer learning, where the model is first pretrained on the source domain and then fine-tuned on the target domain. The results are compared with simple baseline models, including a lagged-target naive model and a regional mean benchmark.

The findings show that transfer learning substantially improves predictive accuracy. Compared with the target-only CatBoost model, the transfer learning model reduces RMSE from 4.0421 to 3.0426 and MAE from 3.3504 to 2.3369. also improves from − 5.8205 to − 0.0761. These results suggest that transfer learning can help stabilize machine learning forecasts when target-domain observations are limited.

Lung Disease Classification Using Transfer Learning on Chest X-ray Images

/in /by

Lung diseases remain a significant global health concern, necessitating the development of rapid and accurate diagnostic methods. While previous research has shown the promise of deep learning models, particularly transfer learning with architectures such as ResNet and VGG, limitations persist in evaluation scope, class imbalance handling, and model interpretability. This study proposes an enhanced deep learning framework for multi-label classification of thoracic diseases using chest X-ray images, addressing these gaps through comprehensive evaluation metrics, advanced data augmentation, and explainable AI (XAI) techniques. The NIH ChestX-ray14 dataset is utilized, with class imbalance mitigated via synthetic minority oversampling and weighted focal loss. Multiple state-of-the-art CNN architectures, including EfficientNet and ResNet variants, are benchmarked using precision, recall, F1 Score, AUC, and accuracy. Moreover, Gradient-weighted Class Activation Mapping (Grad-CAM) is integrated to visualize pathological regions, improving clinical interpretability. The offered framework can perform better in all assessment criteria, achieving an AUC of 0.91 with EfficientNet-B0, and provides interpretable outputs critical for deployment in real-world diagnostic settings. This work advances automated radiological diagnosis by addressing key methodological shortcomings and offers a reliable, explainable solution for lung disease detection.

Detection of COVID-19 using Modified VGG Architectures

/in /by

COVID-19 has created havoc in the world. This paper aims to study and understand the performance of modified VGG-16 and VGG-19 architectures in detecting COVID-19 using the concept of transfer learning. The algorithm has been validated using a private dataset with normal and COVID-19 positive chest X-ray images.

COVID-19 has created havoc in the world. This paper aims to study and understand the performance of modified VGG-16 and VGG-19 architectures in detecting COVID-19 using the concept of transfer learning. The algorithm has been validated using a private dataset with normal and COVID-19 positive chest X-ray images.