Wisdom of a few

Zheng @ playpoly Sr 20091105

I. Cold Start

Greg Linden's latest paper: "The Wisdom of the few: a collaborative filtering approach based on expert opinions from the Web" (PDF, the wisdom of a few people: collaborative Filtering based on the opinions of network experts) made the following comments:

"

What they do say is that using a very small pool of experts works surprisingly well.

In this paper, with a small pool of experts, the recommendation results are amazing.

In particle, I think it suggests a good alternative to content-based methods for bootstrapping a Recommender System.

I think it provides a good alternative for a recommendation system self-launch.

If you can create a high quality pool of experts, even a fairly small one, you may have good results starting with that while you work to gather ratings from the broader community.

"

That is, you can select a high-quality expert pool, which can be your team or the selected expert group, even a relatively small group, your recommendation system will also have a very good start. The wisdom of a few people can solve the Cold Start Problem of the recommendation system at this moment. This is also the reason why playgroup Sr first chose Experts Pool as its origin, and there was a good information filter effect.

 

Ii. Abstract:

For ease of understanding, I will share this paper below:

Nearest-neighbor collaborative filtering is an effective recommendation method. But it is always suffering from these problems:

Data sparsity and noise; cold start (cold-start); scalability.

Therefore, the author proposes a new method, a variant of the traditional collaborative filtering method:

The nearest neighbor is not implemented for user-rating data.AlgorithmInstead, an expert neighbors set is used as a comparison sample to calculate the similarity between this group of people and target users.

This method does not have much scalability problems at least, which is equivalent to narrowing down the benchmark set for comparison. The original method of Nearest Neighbor can be roughly understood as two-to-two comparison. Computing takes some time, and when new users (especially the arrival of a certain tourist group will make the Data Noise more messy) are everywhere, few pieces of data allow you to calculate similarity.

 

The author defines experts as an independent individual who can generate thoughtful, consistent and reliable assessment (scoring) and trust in a given field.

(Original:

We define an expert as an individual that we can
Trust to have produced thoughtful, consistent and reliable
Evaluations (ratings) of items in a given domain.

)

 

We pay more attention to the following two perspectives of the authors:

(A) Study how preferences of a large population can be pre-
Dicted by using a very small set of users;

The study uses a small group of users to predict the value of massive users;

(C) analyze whether professional raters are good predictors for general users;

 

If these aspects are feasible, you do not need to obtain a massive number of users.CommunityAll the data of, as long as the Experts Pool is locked, can be recommended for users.

 

Appendix:

The original text of Greg Linden in the blocked Blogspot is as follows:

Wednesday, November 04,200 9

Using only experts for recommendations

A recent paper from SIGIR, "the wisdom of the few: a collaborative filtering approach based on expert opinions from the Web" (PDF ), has a very useful authentication into the specified tiveness of recommendations using only a small pool of trusted experts.
The results suggest that using a small pool of a couple hundred experts, possibly your own experts or experts selected and mined from the web, has quite a bit of value, especially in cases where big data from a large community is unavailable.
A brief excerpt from the paper:

recommending items to users based on expert opinions .... addresses some of the specified comings of traditional Cf: data sparsity, scalability, noise in user feedback, privacy, and the cold-start problem .... [our] method's performance is comparable to traditional CF algorithms, even when using an extremely small expert set .... [of] 169 experts.
our approach requires obtaining a set... experts... [we] crawled the rotten tomatoes Web site -- which aggregates the opinions of movie critics from various media sources -- to obtain expert ratings of the movies in the Netflix data set.

The authors certainly do not claim that using a small pool of experts is better than traditional collaborative filtering.
What they do say is that using a very small pool of experts works surprisingly well. In particle, I think it suggests a good alternative to content-based methods for bootstrapping a Recommender System.If you can create a high quality pool of experts, even a fairly small one, you may have good results starting with that while you work to gather ratings from the broader community.

 

OthersArticle:

Just remember to learn about the journey of searching for ghosts

Do not trust pubsubhubbub too much

Just remember to know the series

Semantics and features