Inferring Multi-Dimensional Ideal Points for US Supreme Court Justices
Authors: Mohammad Islam, K. S. M. Hossain, Siddharth Krishnan, Naren Ramakrishnan
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate SCIPM using both synthetic and real-world datasets. Our experiments are designed to model how justices voted for specific opinions, to understand the quality of inferred ideal points for justices as well as for opinions and whether the inferred points can capture dynamics of coalitions. |
| Researcher Affiliation | Academia | Mohammad Raihanul Islam , K.S.M. Tozammel Hossain , Siddharth Krishnan, and Naren Ramakrishnan Discovery Analytics Center, Department of Computer Science, Virginia Tech Email: {raihan8, tozammel, siddkris, naren}@cs.vt.edu |
| Pseudocode | Yes | Algorithm 1 Op Gen(β, D, U, K, ϵ, Nmax, χ, μx, σx, μa, σa) |
| Open Source Code | Yes | Code and Data for the experiments are available online (see Supplementary Material (Islam et al. 2016)). |
| Open Datasets | Yes | We collected opinion files3 for the current Court (from 2010 to 2014) and curated them into a structured data format. |
| Dataset Splits | Yes | We predict justices who vote for an opinion using a 5-fold cross-validation over both real-world and synthetic datasets (see Sec. 5). We report the average recall value over five test-folds for various parameter settings. |
| Hardware Specification | No | The paper does not mention any specific hardware used for running experiments, such as CPU or GPU models, or memory specifications. |
| Software Dependencies | No | The paper mentions using algorithms like LDA and Naive Bayes but does not specify any software libraries or tools with version numbers for their implementation. Although the supplementary material mentions MATLAB, no version number is provided in the main text. |
| Experiment Setup | Yes | For both unsupervised and supervised settings, we use μa = 0 and σa = 0.1 as priors for opinions ideal points... With an unbiased prior for justices ideal points... we set μx = 0 and σx = 1. With a biased prior... we set σx = 0.5 with μx = 1 for Republicans and μx = 1 for Democrats. |