Quiz 3 - Data Mining Quiz 3 Recommender Systems The Long...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Data Mining – Quiz 3 Recommender Systems The Long Tail: a substantial portion of the demand pertains to a large number of items which have relatively very little demand What product categories are likely to have a long-tailed demand? A large number of alternatives Profiting from the long tail: Challenges for consumer in exploring products of potential interest in the long tail? RECOMMENDATIONS Mass customization: Recommender systems generate personalized recommendations for products or services - Monetize demand for niche items – give legal value/put a dollar amount on - Enhance demand for niche products opportunities for secondary markets TRADITIONAL CONTENT-BASED modeling approach Ex: Pandora, BookLamp Assume customers rate product on a scale of 1-5 For each customer, construct a data set of the products the customer rated - For each product: the independent variables ( predictors) are the product attributes the rating is the target Induce a model from the data and use the model to predict the user’s ratings for other restaurants and identify the best ones to recommend ** Come up with good predictors Each instance is a product. Induce model mapping product attributes to ratings. Drawbacks: When preferences are complex (difficult to describe/quantify) or entirely tacit (personal taste and judgment) - For some customers, important predictors (drivers) of the customer’s preferences may be tacit (not understood even by the customer him/herself) - Some predictors may be known but difficult to quantify - Some drivers of preference are idiosyncratic to the product and cannot be defined generally for all products. Ex: liking a restaurant for a single item on the menu regardless of other, potentially undesirable, properties - For each item, extracting the value for each predictor may require significant resources Potentially important predictors of customers’ preferences may not be included in the model
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.
  • Spring '11
  • Saar-Tsechansky
  • K-nearest neighbor algorithm, Nearest neighbor search, Collaborative filtering, Statistical classification, nearest neighbors

{[ snackBarMessage ]}

Page1 / 2

Quiz 3 - Data Mining Quiz 3 Recommender Systems The Long...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online