Reading Thomas Jefferson with TopicViz: Towards a Thematic Method for Exploring Large Cultural Archives
Reading Thomas Jefferson with TopicViz: Towards a Thematic Method for Exploring Large Cultural Archives
复制标题
用 TopicViz 阅读托马斯·杰斐逊：探索大型文化档案的主题方法
DOI：
10.22230/src.2013v4n3a121
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
Jacob Eisenstein
中科院分区：
文献类型：
--
作者：
Lauren F. Klein;Jacob Eisenstein
0 0 1 144 823纽约城市大学6 1 966 14.0正常0 false false EN-US JA X-NONE/* 样式定义 */
table.MsoNormalTable
{mso-style-name：“表格正常”;
mso-tstyle-rowband-size：0;
mso-t样式colband大小：0;
mso样式不显示：是;
mso-style-priority：99;
mso样式父代：“";
mso-填充-alt：0英寸5.4磅0英寸5.4磅;
mso-para-margin：0英寸;
mso-para-margin-bottom：.0001pt;
mso-pagination：鳏夫孤儿;
font-size：10.0pt;
font-family：“Times New Roman”;}
尽管艾德福尔松称之为“档案的史诗般转变”，指的是从印刷到数字档案形式的转变，但探索这些数字化收藏的方法仍然不发达。数字化所促进的一种方法是应用自动文本挖掘技术，如主题建模--一种用于识别文档档案中重复出现的主题的计算方法。我们回顾了文学档案主题建模的新兴文献，并提出了一个案例研究，应用主题模型的文件托马斯杰斐逊。这项工作的经验教训表明，前进的方向是为学者提供更全面的支持，以可视化和探索主题模型输出，同时将主题模型与更传统的工作流程集成，这些工作流程围绕组装和完善相关文档集。我们描述了我们正在努力开发一个新的软件系统，实现这些想法。
0 0 1 144 823 City University of New York 6 1 966 14.0 Normal 0 false false false EN-US JA X-NONE /* Style Definitions */
table.MsoNormalTable
{mso-style-name:"Table Normal";
mso-tstyle-rowband-size:0;
mso-tstyle-colband-size:0;
mso-style-noshow:yes;
mso-style-priority:99;
mso-style-parent:"";
mso-padding-alt:0in 5.4pt 0in 5.4pt;
mso-para-margin:0in;
mso-para-margin-bottom:.0001pt;
mso-pagination:widow-orphan;
font-size:10.0pt;
font-family:"Times New Roman";}
In spite of what Ed Folsom has called the “epic transformation of archives,” referring to the shift from print to digital archival form, methods for exploring these digitized collections remain underdeveloped. One method prompted by digitization is the application of automated text mining techniques such as topic modeling -- a computational method for identifying the themes that recur across an archive of documents. We review the nascent literature on topic modeling of literary archives, and present a case study, applying a topic model to the Papers of Thomas Jefferson. The lessons from this work suggest that the way forward is to provide scholars with more holistic support for visualization and exploration of topic model output, while integrating topic models with more traditional workflows oriented around assembling and refining sets of relevant documents. We describe our ongoing effort to develop a novel software system that implements these ideas.