Generalized Model Aggregation

The Cradle via Flickr

Generalized Model Aggregation (GMA)

A five-year joint project between MIT Sloan School of Management and the School of Industrial & Systems Engineering at Georgia Tech

Rapid growth in scientific output requires flexible and robust methods for aggregating the findings from prior studies. In most cases, 'qualitative' reviews are conducted for taking stock of what is known, but they offer little 'quantitative' guidance. Also, the common quantitative aggregation methods (i.e., meta-analysis methods) usually combine one explanatory variable (e.g., a treatment) on one response variable (e.g., a health outcome) across multiple studies with similar designs. 

Despite these limitations, the rapid growth of scientific literature has notably promoted increasing applications of meta-analysis. The number of articles in major databases with the term “meta-analysis” ONLY in the title shows over 25-fold growth over the last decade—reaching tens of thousands annually. Therefore, the value of a broader and more flexible method for synthesizing prior research can be immense across various disciplines.

Our recent paper in PLOS ONE outlines a new method (generalized model aggregation – GMA; see it on Wikipedia!) for aggregating into a meta-model the results of prior studies (of a phenomenon) when those studies vary in design and measures used. In the paper, we provide several numerical and empirical examples demonstrating the ability of GMA to aggregate evidence from methodologically diverse studies and obtain unbiased estimates from potentially mis-specified studies. By enabling more complex meta-analyses, GMA allows researchers to leverage previous findings to compare alternative theories and advance new models in diverse domains.


      Semiars and presentations


    Oct 13, 2017, time: TBA, Zeppelin University, Germany (location: TBA), by Dr. Jalali

    Sep 20, 2017, 12 to 1:30 pm, Harvard IQSS, (location: TBA), by Dr. Jalali

    July 3, 2017, 2:30 pm, McGill University, (SRSM Annula Meeting in Trottier Building), by Dr. Jalali

    May 5, 2017, 2 pm to 3:30 pm, Harvard School of Public Health (building 2, Room 426), by Dr. Jalali 

    Apr 2017, Rand Corporation, by Dr. Rahmandad

    March 2017, University of Tokyo, by Dr. Jalali 

    Feb 2017, Tokyo Institute of Technology, by Dr. Jalali 


         Media Coverage

        How to boil down a pile of diverse research papers into one cohesive picture, May 2017

        MIT Sloan Professor Builds New Meta-analysis Method to Help Settle Unresolved Debates, April 2017

        New Meta-Analysis Method To Help Settle Unresolved Debates, April 2017

        Why Meta-Data Analysis Tools Like This one are key for Industry 4.0, April 2017



        See/downlaod the paper PLOS ONE
        Downlaod supplementary text
        See below for GMA codes and instructions.


         Potential applications of GMA 

        Here are some articles that seem to be potentially relevant applications for GMA. They often cunducted a systematic review of the literature and listed prior studies, but did not aggregate the findings quantitatively. Please note that these are found using a quick search on the Internet. There are many more problems that can benefit from GMA. If you are interested in applying GMA for any problem and need our assitance, please shoot us an email!

        Health: Prediction models for colorectal cancer (link)  |  Prediction models for lung cancer (link)  |  Models for cardiovascular disease risk (link)  |  Risk models for diabetes complications (link)  

        Social sciences: Influence of climate on human conflict (link)

        Management: Sustainable supply chain management (link)  |  Supplier selection (link)


         GMA on WIKIPEDIA

        Generalized Model Aggregation


         Recipe for GMA and Codes

        Here we provide a quick recipe of the main steps to use GMA. Check out the supplementary document for the details of each step. Also, the Codes README file provides more information about the code files listed in each step--as well as all other code files, including their actions, inputs, and outputs.  



        GMA Codes (Download MATLAB codes, Version 2)

        Note: To run the MATLAB codes, open GMA_RUN.m in MATAB Editor and execute the codes. The current codes present Scenario 1 in the paper.
        See more detailed instructions in the README document.

        • README for MATLAB Codes
        • Questions? Feel free to send an email to:
        • The same codes (Version 2) were shared as a supplementary file for the PLOS ONE article (2017). 

        GMA Codes (Downlaod R codes, Version 1 Beta)

        See the instructions for the codes in the README document. 

        • README for R Codes
        • Questions? Feel free to send an email to:
        • Credit: These codes are converted from MATLAB and adjusted for R by Rebecca Salzman.


        Cite these materials as:

        Rahmandad H, Jalali MS, Paynabar K (2017) A flexible method for aggregation of prior statistical findings. PLOS ONE 12(4): e0175111. doi: 10.1371/journal.pone.0175111

        Dowload the citation here