Estimating the Distribution of Displacements

TL;DR

This paper introduces a simple least squares-based method for estimating the distribution of time differences between connected events without needing to match the events directly, simplifying analysis in various applications.

Contribution

The paper presents a novel, straightforward approach for distribution estimation that does not require event matching, unlike traditional methods.

Findings

Effective estimation of time difference distributions without event matching

Method's simplicity facilitates easy implementation

Applicable to various connected event analysis scenarios

Abstract

In this paper, we propose a method for estimating the distribution of time differences between connected events (such as ad impressions and corresponding customer calls). A special feature of this method is that it does not require matching those connected events with each other. The method is very simple to use as it essentially consists of computing an ordinary least squares estimator.