Files in this item

FilesDescriptionFormat

application/pdf

application/pdfZhu, Junzhe.pdf (229kB)
PosterPDF

application/pdf

application/pdfZhu, Junzhe.pdf (288kB)
ImagePDF

Description

Title:A Tree in a Random Forest
Author(s):Zhu, Junzhe
Contributor(s):Wickes, Elizabeth; Chuang, Paul
Subject(s):Electrical Engineering
Topic Modelling
Decision Forest
Abstract:We web-scraped 450000+ comments from New York Times, applied a LDA model to analyze their topic distribution, and manually marked these topics. We then derived a Random Decision Tree Forest of 100 regression trees, with more than 1100 nodes in each tree, in order to infer audience engagement (indicated by comment length) from topic distribution for each comment. The visualization of the tree helped us determine the importance of each topic in terms of how it affects audience's attention.
Issue Date:2019
Type:Image
URI:http://hdl.handle.net/2142/103810
Rights Information:Copyright 2019 Junzhe Zhu
Date Available in IDEALS:2019-05-10


This item appears in the following Collection(s)

Item Statistics