P2P Lending Analysis Using the Most Relevant Graph-Based Features

Abstract

Abstract. Peer-to-Peer (P2P) lending is an online platform to facilitate borrowing and investment transactions. A central problem for these P2P platforms is how to identify the most influential factors that are closely related to the credit risks. This problem is inherently complex due to the various forms of risks and the numerous influencing factors involved. Moreover, raw data of P2P lending are often high-dimension, highly correlated and unstable, making the problem more untractable by traditional statistical and machine learning approaches. To address these problems, we develop a novel filter-based feature selection method for P2P lending analysis. Unlike most traditional feature selection methods that use vectorial features, the proposed method is based on graph-based features and thus incorporates the relationships between pairwise feature samples into the feature selection process. Since the graph-based features are by nature completed weighted graphs, we use the steady state random walk to encapsulate the main characteristics of the graph-based features. Specifically, we compute a probability distribution of the walk visiting the vertices. Furthermore, we measure the discriminant power of each graph-based feature with respect to the target feature, through the Jensen-Shannon divergence measure between the probability distributions from the random walks. We select an optimal subset of features based on the most relevant graph-based features, through the Jensen-Shannon divergence measure. Unlike most existing state-of-the-art feature selection methods, the proposed method can accommodate both continuous and discrete target features. Experiments demonstrate the effectiveness and usefulness of the proposed feature selection algorithm on the problem of P2P lending platforms in China.

Identifying the most informative features using a structurally interacting elastic net

“…This deficiency restricts the precision of the information theoretic measure between pairs of features. To address this drawback, Cui et al [26] have recently developed a new feature selection method using graph-based features. Specifically, they transform each feature vector into a graph structure that encapsulates pairwise relationship between samples.…”

Adaptive Feature Selection Based on the Most Informative Graph-Based Features

“…To further evaluate our study, we compare the proposed method (AFS-QW) with several alternative feature selection methods. These alternative methods include: 1) the Fisher Score method (FS) [13], 2) the Mutual Information based method (MI) [19], and 3) most relevant graph-based feature selection method (FS-RW) [11]. The classification accuracy of each method is shown in Fig.…”

“…These include correlation analysis (CA) and multiple linear regression (ML-R), which are simple but widely applied. Furthermore, we also compare the proposed method to the most relevant graph-based feature selection method associated with the SSRW (FS-RW) [11], since it can also accommodate the continuous target feature. Table 1 presents a comparison of the results obtained using these methods.…”

“…This drawback limits the precise information theoretic measure between pairwise features. To address this shortcoming, Cui et al [11] have proposed a new feature selection method in terms of graph-based features. They transform each vectorial feature into a graph structure that encapsulates the relationship between pairwise samples from the feature.…”

“…In summary, the main contributions are threefold. First, like Cui et al [11], for the above dataset X having N features, we transform each vectorial feature f i into a graph-based feature G i . Here, G i is a complete weighted graph, where each vertex v a represents a corresponding sample f ia in f i and each weighted edge {v a , v b } represents the relationship between pairwise samples f ia and f ib .…”