Mohammad Javad Nadjafi Arani

Fraudulent claims in the automobile industry pose a significant threat to the financial stability of insurance companies and erode the trust between policyholders and insurers. Organized fraud, which involves intricate schemes and multiple parties, presents a substantial challenge in detection due to imbalanced datasets. While existing techniques such as over-sampling and under-sampling have been proposed to address this issue, they often lead to overfitting, loss of information, and reduced accuracy. However, assigning a suspicious label to each policyholder is more changeable, as it can identify potential risks and prevent fraudulent activities before they occur. In response to these challenges, we propose a novel heuristic approach called Organized Fraud detection with Graph theory and Poisson process (OrFGP) that identifies suspiciously organized fraud groups within an accident network and provides credibility levels for accidents and associated individuals. We first demonstrate that car accidents follow a Poisson random process. We then combine this process with graph theory to introduce an accident network. In the network, our objective is to identify regular behavior between accidents, which, based on the stochastic nature of accidents, can indicate organized fraud. OrFGP uses probabilistic concepts in conjunction with local network connectivity metrics to evaluate the credibility of accidents and individuals. The results indicate that OrFGP outperforms state-of-the-art approaches, particularly in imbalanced datasets. In fact, OrFGP achieves an accuracy of 98% and improves the F1-score by at least 3%.

Quantitative Structure Property Relationship (QSPR) modeling is critical for predicting physicochemical properties of molecules, supporting drug discovery and material design. Traditional methods often use complex descriptors or opaque models, which limits interpretability. There is a need for QSPR approaches that balance accuracy and interpretability to elucidate structural influences on molecular properties. We introduce a QSPR method using degree-distance-based topological indices (TIs) derived from vertex edge weighted (VEW) molecular graphs, weighted by atomic number, mass, radius, density, electronegativity, and ionization energy. This approach captures detailed molecular connectivity and bonding while prioritizing interpretability. Using a 166-molecule dataset, our models -Ridge Regression, Random Forest, XGBoost, and Neural Networks- achieved high prediction accuracy for six physicochemical properties. Regularization ensured robust predictions. The performance metrics tables and the TI correlations clarified the structure-property relationships. This efficient and interpretable framework accelerates drug discovery by enabling virtual screening and informed molecular design.

Combining topology control with network coding has been recently noticed which provides essential advantages such as lower energy consumption and higher lifetime. In this paper, we utilize optimization and graph theory to propose a mathematical perspective for combining network coding with transmission range adjustment in Wireless Sensor Network (WSN). For this aim, we define an extended optimization problem based on two vectors of variables including transmission range and network flow. In particular, it is proved that there is a coding solution if and only if the extended optimization problem has at least an optimal solution. Due to this theorem, a new objective function is provided which is more practical to prolong the lifetime of WSNs. The simulation results demonstrate that the proposed model could adjust transmission range efficiently to obtain the optimal lifetime for network-coding-based WSNs.