Data Mining and Text Analytics Advertising Laura Quinn Internet-based Advertising The internet-based advertising industry is worth billions and has grown massively in recent years. It is constantly expanding and therefore requires robust systems which are able to handle billions of queries from users in order to provide ads from millions of advertisers on many websites. This leads to a massive reliance on: * Real-time collection of data * Management and analysis of the data in order to provide appropriate ads. http://www.dummies.com/how-to/content/text-analytics-for-unstructured-big-data.html Internet-based Advertising These advertisements exist in many forms including: * Sponsored Search * Display Advertising * Rich Media Ads * Interstitial Ads Google Sponsored Search Advertisers pay Google only when their ad is clicked. The ads may be shown when a user query contains keywords matching the websites and ads keywords, this helps provide ads tailored to the users search. Collection of Data from Google We may show you ads based on many factors, including: Types of websites you visit, and mobile apps you have on your device The DoubleClick cookie on your browser and the settings in your Ads Settings Websites and apps you’ve visited that belong to businesses that advertise with Google Previous interactions with Google’s ads or advertising services Your Google or YouTube profile We don’t do the following: Link your name or personally identifiable information to your DoubleClick cookie without your consent Associate your DoubleClick cookie with sensitive topics like race, religion, sexual orientation, or health without your consent Google and YouTube Internet-based Advertising Google Ads Settings based on websites that you have visited. Without logging in to Google, it can determine users: * Gender * Age * Languages * Interests Data Mining and Text Analytics Algorithms Off-line algorithms are used by observing data over a period of time and make decisions from the data . In this case, this would be looking at search queries from the time period and the advertisers bids for those queries. However, this would not be useful for the sponsored search example as it would not provide realtime results. Data Mining and Text Analytics Algorithms On-line algorithms are a better way to show sponsored search ads as they do not require the entire data input from the beginning. They are able to use information about: * The advertisers remaining budget * The click-through rate These can be classed as a greedy algorithm, this means that they try to find the best solution after each input but this does not necessarily mean it will be the optimum solution in terms of the profit for the advertiser and also the search engine. Advertising Algorithms In Conclusion The online advertising industry must continue to tackle many new and existing challenges. These include how to understand the needs of the users and also what the advertiser wants to achieve. It must also deal with which type of advertising is the most suitable for each of the different advertisements, as well as the challenges faced with continuing to provide the user with customized and useful ads. These are challenges which will need to be tackled by data miners and ad companies to find new technologies and systems.