Graduation Year

2015

Date of Submission

4-2015

Document Type

Open Access Senior Thesis

Degree Name

Bachelor of Arts

Reader 1

Arthur Lee

Reader 2

America Chambers

Terms of Use & License Information

Terms of Use for work posted in Scholarship@Claremont.

Rights Information

© 2015 Erin Chavanne

Abstract

Services that once required human interaction are now completed with the click of a few buttons. In general, this allows for a more streamlined process for activities such as sending messages (email or text messages), filing taxes, or even shopping for groceries. In terms of searching for hotels and travel accommodations however, this process has not proven to be the most effective as the speed and efficiency is hindered by the interface through which this information is available. Choosing a travel specific site, filling in the required fields, combing through results for the desired specifications, and then possibly repeating the process elsewhere, does not provide the ability for the user to express the entirety of their preferences for the accommodation and is therefore not an effective method for searching.

Natural language search provides a more accessible and intuitive interface for accommodation searching. Instead of specifying fields that may not encompass the the entirety of the desired search, the user is able to express all of the aspects in a single, natural language, search.

In this project, we propose a natural language search interface for accommodations such as hotels, hostels, or apartments. Data acquired through Amazon Mechanical Turk is used to create a system for extracting various accommodation fields. Zilyo and Expedia APIs are then queried for real-time accommodation listings. These results are then adjusted based on the specifics of the search that were not included in the original query. A natural language search of this kind is not only more accessible on the consumer end, but provides data that pertains directly to the the entirety of the intended search.

Download

Share

COinS