Browse Prior Art Database

Cognitive Site Creation

IP.com Disclosure Number: IPCOM000250115D
Publication Date: 2017-Jun-01
Document File: 6 page(s) / 367K

Publishing Venue

The IP.com Prior Art Database

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 37% of the total text.

TITLE: Cognitive Site Creation

Abstract

This disclosure focuses on providing a cognitive recommender framework, which connects the analysis and the recommendation phase of web site creation. Criteria for measuring success are based on existing sentiments available in the web. Those reflect end-user’s positive and negative opinions which have been made about literally all existing web sites in the Internet. The acquired knowledge is presented to web site authors.

Background

Today, website authors are building new websites based on their experience and intuition what they hope works best for the end users. While it is essential to target web site appearance and content to the desired consumers, it is hard to obtain the necessary guidance and recommendations during the time a web site is being designed. This disclosure utilizes a combination of sentiment analysis based on successful web sites as well as cognitive capabilities to give clear recommendations to a web site author.

Prior Art

Website authors are able to use tools like A/B or multivariate testing to get feedback over time on how different variations of the newly created websites are resonating with customers. Behavioral Analytics Data can also be gathered from actual usage patterns. Typically, that historic data is used to propose “next best actions” to the web site author. Recommendations can be grouped to targeting segments to provide similar experiences to similar end-user personalities. Nevertheless, all those methods can be used to provide the website author with very valuable data and recommendations to tune and improve the website, after the site is live. In addition, it is significant work to identify user segments and rules to configure the analytics systems. They need significant configuration and manual training (e.g. defining meaningful user segments …). Such segments and rules are very specific to the business context and unlikely to be generic enough to be shared across industries. Analytics capabilities are also strongly focused on the feedback on the own web site, rather than offering a cross-industry picture. Simple recommendations like they are done in Facebook and other consumer-facing authoring environments are strongly tied to very limited use cases.

Summary – Part 1

This disclosure focuses on providing a framework, which connects the analysis and the recommendation phase of web site creation. Criteria for measuring success are based

on existing sentiments available in the web. Those reflect end-user’s positive and negative opinions which have been made about literally all existing web sites in the Internet (as long as public sentiments can be found during the sentiment analysis). Sentiments are condensed statements describing a characteristic tag/criteria of an opinion about a web site, e.g. “hard to comprehend”, “many images”, … . The Sentiments are retrieved in the initial data mining phase of the system. The sentiment retrieval includes an automatic categorizat...