InfoTrad: An R package for estimating the probability of informed trading

The purpose of this paper is to introduce the R package InfoTrad for estimating the probability of informed trading (PIN) initially proposed by Easley et al. (1996). PIN is a popular information asymmetry measure that proxies the proportion of informed traders in the market. This study provides a short survey on alternative estimation techniques for the PIN. There are many problems documented in the existing literature in estimating PIN. InfoTrad package aims to address two problems. First, the sequential trading structure proposed by Easley et al. (1996) and later extended by Easley et al. (2002) is prone to sample selection bias for stocks with large trading volumes, due to floating point exception. This problem is solved by different factorizations provided by Easley et al. (2010) (EHO factorization) and Lin and Ke (2011) (LK factorization). Second, the estimates are prone to bias due to boundary solutions. A grid-search algorithm (YZ algorithm) is proposed by Yan and Zhang (2012) to overcome the bias introduced due to boundary estimates. In recent years, clustering algorithms have become popular due to their flexibility in quickly handling large data sets. Gan et al. (2015) propose an algorithm (GAN algorithm) to estimate PIN using hierarchical agglomerative clustering which is later extended by Ersan and A (2016) (EA algorithm). The package InfoTrad offers LK and EHO factorizations given an input matrix and initial parameter vector. In addition, these factorizations can be used to estimate PIN through YZ algorithm, GAN algorithm and EA algorithm.