How to Cautiously Uncover the “Black Box” of Machine Learning Models for Legislative Scholars

Machine learning models, especially ensemble and tree-based approaches, offer great promise to legislative scholars. However, they are heavily underutilized outside of narrow applications to text and networks. We believe this is because they are difficult to interpret: while the models are extremely flexible, they have been criticized as “black box” techniques due to their difficulty in visualizing the effect of predictors on the outcome of interest. In order to make these models more useful for legislative scholars, we introduce a framework integrating machine learning models with traditional parametric approaches. We then review three interpretative plotting strategies that scholars can use to bring a substantive interpretation to their machine learning models. For each, we explain the plotting strategy, when to use it, and how to interpret it. We then put these plots in action by revisiting two recent articles from Legislative Studies Quarterly.     

Can the RUVIS Reflected UV Imaging System Visualize Fingerprint Corrosion on Brass Cartridge Casings Postfiring?

Comparisons are made between the visualization of fingerprint corrosion ridge detail on fired brass cartridge casings, where fingerprint sweat was deposited prefiring, using both ultraviolet (UV) and visible (natural daylight) light sources. A reflected ultraviolet imaging system (RUVIS), normally used for visualizing latent fingerprint sweat deposits, is compared with optical interference and digital color mapping of visible light, the latter using apparatus constructed to easily enable selection of the optimum viewing angle. Results show that reflected UV, with a monochromatic UV source of 254 nm, was unable to visualize fingerprint ridge detail on any of 12 casings analyzed, whereas optical interference and digital color mapping using natural daylight yielded ridge detail on three casings. Reasons for the lack of success with RUVIS are discussed in terms of the variation in thickness of the thin film of metal oxide corrosion and absorption wavelengths for the corrosion products of brass.     

Visualizing Digital Forensic Datasets: A Proof of Concept

Digital forensic visualization is an understudied area despite its potential to achieve significant improvements in the efficiency of an investigation, criminal or civil. In this study, a three‐stage forensic data storage and visualization life cycle is presented. The first stage is the decoding of data, which involves preparing both structured and unstructured data for storage. In the storage stage, data are stored within our proposed database schema designed for ensuring data integrity and speed of storage and retrieval. The final stage is the visualization of stored data in a manner that facilitates user interaction. These functionalities are implemented in a proof of concept to demonstrate the utility of the proposed life cycle. The proof of concept demonstrates the utility of the proposed approach for the storage and visualization of digital forensic data.     

Changes in Fingerprints Depending on Physiological Factors

The aim of the study was to determine the effect of gender, body mass index, and time on the width of friction ridge impressions in fingerprints and the content of sweat and sebum forming the print. The research showed that the friction ridge impressions of women are significantly wider than those of men, and this tendency persists throughout the existence of the print. The differences may result from the anatomical structure of the skin, as well as from the use of protective products such as creams, which are used more often by women. The width of friction ridge impression is similar in underweight and overweight people, but greater than in those with normal body weight, and the amount of sweat and sebum increases with body weight. The passage of time significantly reduces both the width of the friction ridge impression and the content of sweat and sebum secretions in the print.     

Enhanced Visualization of Latent Fingermarks on Rough Aluminum Surfaces Using Sequential Au and Zn/ZnS/ZnO Depositions

Detection and visualization of fingermarks on rough and diffuse surfaces is a relatively challenging task. We succeeded in developing latent fingermarks on scratched and rough aluminum surfaces by sequential deposition of a thin layer of gold followed by one of zinc or zinc‐based compounds on the fingermarks. The best image enhancement was achieved with sequential Au and ZnS depositions. Using this combination, we could enhance the visualization of latent fingermarks aged over 65 days in normal conditions. The optical reflectance from the fingermarks with the deposited layers of metal/dielectric is analyzed as a stratified medium. Significant contrast in the reflectance from the regions of the ridges and the valleys of the fingermark would enhance the visualization. Our results show that the Au and ZnS bi‐layer combination can have a large reflection contrast and improved fingermark visualization at wavelengths corresponding to the green light for specific thickness of ZnS.     

Visualization of latent fingerprints using fluorescence lifetime imaging on paper emitting strong fluorescence

Latent fingerprints were successfully visualized using fluorescence lifetime imaging (FLIM) on paper which emits strong fluorescence with a lifetime close to that of fingerprints and thus from which it is difficult for time-resolved spectroscopy to visualize fingerprints. Latent fingerprint samples on paper were excited using a 450 nm or 532 nm nanosecond pulsed-laser, and time-resolved fluorescence images were obtained at a delay time of 6–16 ns in intervals of 1 ns, to the excitation pulse. The excitation beam was expanded using a lens, and the fluorescence from the fingerprints was captured using an intensified CCD camera. Because of the large fluorescence intensity of the background paper of approximately two to four orders of magnitude larger than that of the fingerprint, the fingerprint was not visualized on each fluorescence image by time-resolved spectroscopy. However, the fingerprint was visualized in a FLIM image constructed using a series of the fluorescence images for the case with the fluorescence intensity of the background paper being four orders of magnitude larger than that of the fingerprint. The difference in fluorescence lifetime in the FLIM image of the visualized fingerprint and background paper was in the order of 0.1 ns, which was an order of magnitude smaller than the inherent fluorescence lifetime of a few nanoseconds for the fingerprints and paper. It was demonstrated that, at a background fluorescence intensity with a certain order of magnitude larger than that of fingerprints, FLIM has the potential to visualize latent fingerprints which cannot be visualized by time-resolved spectroscopy.     

Rapid Localization of Bone Fragments on Surfaces using Back‐Projection and Hyperspectral Imaging

Manual localization of bone fragments on the ground or on complex surfaces in relation to accidents or criminal activity may be time‐consuming and challenging. It is here investigated whether combining a near‐infrared hyperspectral camera and chemometric modeling with false color back‐projection can be used for rapid localization of bone fragments. The approach is noninvasive and highlights the spatial distribution of various compounds/properties to facilitate manual inspection of surfaces. Discriminant partial least squares regression is used to classify between bone and nonbone spectra from the hyperspectral camera. A predictive model (>95% prediction ability) is constructed from raw chicken bones mixed with stone, sand, leaves, moss, and wood. The model uses features in the near‐infrared spectrum which may be selective for bones in general and is able to identify a wide variety of bones from different animals and contexts, including aged and weathered bone.     

党组织嵌入城市社区治理的逻辑框架构建——基于Cite space的文献计量分析

为了解城市社区治理研究的文献特征与发展趋势,借助可视化软件Cite SpaceⅤ,梳理2012年以来CNKI中社区治理研究文献,发现城市社区治理研究集中于东中部地区高校,研究视角呈现多元化和本土化特征,研究内容经历了从管理到治理的蜕变,社区党组织通过整合政府行政力量和社会自治力量在社区治理中发挥了重要作用。据此提出构建结构—过程视角下党组织嵌入城市社区治理的逻辑分析框架,旨在为新时代党组织嵌入城市社区治理研究提供新的思路。     

Whitelisting system state in windows forensic memory visualizations

Examiners in the field of digital forensics regularly encounter enormous amounts of data and must identify the few artifacts of evidentiary value. One challenge these examiners face is manual reconstruction of complex datasets with both hierarchical and associative relationships. The complexity of this data requires significant knowledge, training, and experience to correctly and efficiently examine. Current methods provide text-based representations or low-level visualizations, but levee the task of maintaining global context of system state on the examiner. This research presents a visualization tool that improves analysis methods through simultaneous representation of the hierarchical and associative relationships and local detailed data within a single page application. A novel whitelisting feature further improves analysis by eliminating items of less interest from view. Results from a pilot study demonstrate that the visualization tool can assist examiners to more accurately and quickly identify artifacts of interest.     

A comprehensive micro unmanned aerial vehicle (UAV/Drone) forensic framework

In the early 1990s, unmanned aerial vehicles (UAV) were used exclusively in military applications by various developed countries. Now with its ease of availability and affordability in the electronic device market, this aerial vehicular technology has augmented its familiarity in public and has expanded its usage to countries all over the world. However, expanded use of UAVs, colloquially known as drones, is raising understandable security concerns. With the increasing possibility of drones' misuse and their abilities to get close to critical targets, drones are prone to potentially committing crimes and, therefore, investigation of such activities is a much-needed facet. This motivated us to devise a comprehensive drone forensic framework that includes hardware/physical and digital forensics, proficient enough for the post-flight investigation of drone's activity. For hardware/physical forensics, we propose a model for investigating drone components at the crime scene. Additionally, we propose a robust digital drone forensic application with a primary focus on analyzing the essential log parameters of drones through a graphical user interface (GUI) developed using JavaFX 8.0. This application interface would allow users to extract and examine onboard flight information. It also includes a file converter created for easy and effective 3D flight trajectory visualization. We used two popular drones for conducting this research; namely, DJI Phantom 4 and Yuneec Typhoon H. The interface also provides a visual representation of the sensor recordings from which pieces of evidence could be acquired. Our research is intended to offer the forensic science community a powerful approach for investigating drone-related crimes effectively.