Archive for the ‘recommendations’ Category

Tuesday, September 20th, 2011

If you like this author, how about?

Short version. A few weeks ago, we introduced an “author read-alikes” feature, in four different flavors, with an invitation to help us pick the best one. We got our answer, and have gone with the winner. But the voting worked so well, we’ve decided to integrate it into the new author recommendations. The recommendations—now called “If you like this author..”—all include a ratings option. Rate a recommendation high and it will rise. Rate it low and it will sink. This is fortunate, as our author recommendations—we admit—need work. They’re nowhere near as good as our work-to-work recommendations.

With luck, your input will help us improve them!

More details. The vote between flavors turned out very well for us. Option 2 was a new and extremely slow algorithm. We thought it would be the best, and it polled a respectable 3.43 stars. It certainly outpolled version 3 and 4, at 2.95 and 2.94 stars respectively. But it did slightly worse than option 1, which polled 3.52. Best of all, option 1 was a much easier, faster algorithm. This was unexpected, but welcome. Option one is so fast we’ve been able to apply recommendations to virtually all major authors in a single night—option 2 would have taken weeks or months!

3.52 stars still isn’t that great. But the success of the vote suggested we might do better if we subjected the recommendations themselves to a vote. So we’ve done that. Basically, anything above 3 stars will conspire to move the recommendation up. Anything below three will move it down. The higher or lower the stars, the greater the movement. The effect will differ between authors and recommendation, as all recommendations have a (hidden) score, not just a ranking.

Some examples: Agatha Christie, Sir Arthur Conan Doyle, Umberto Eco, Doris Kearns Goodwin, J. K. Rowling, Malcolm Gladwell.

Come talk about it here.

Labels: recommendations

Tuesday, September 6th, 2011

Recommendations for groups and authors! (but help needed)

I’ve added two new types of “recommendations”—”characteristic works” for member groups and “read-alikes” for author pages. We need your help improving the latter.

Groups recommendations. The group recommendations are on the new and developing “Group Zeitgeist” pages. Each group Zeitgeist includes two lists:

  • Most-held works. Shows the top books held by group members, with no weighting or adjustment–that is, Harry Potter often wins.
  • Characteristic works. Shows the top books, weighted the way recommendations are weighted–that is, it shows works held by group-members in unusual amounts.

“Characteristic works” works quite well. Librarians who LibraryThing lists Taylor’s Introduction to Cataloging and Classification, Library: An Unquiet History and even AACR2. Christianity‘s list starts with Lewis’ Mere Christianity, Cthulhu Mythos with The Cthulhu Mythos Encyclopedia, Medieval Europe with The Civilization of the Middle Ages, etc.

Author read-alikes. The new “author read-alike” uses much the same algorithm, but the results are not always as good. For example, C. S. Lewis recommends George MacDonald, G. K. Chesterton and Madeleine L’Engle–good–but also Laura Ingalls Wilder–read by some of the same people who read Narnia, but not otherwise similar.

To help us improve the algorithm, we’re showing four different versions of the algorithm, and asking members to rate them with stars. Knowing both what authors fail and which version of the algorithm is better will help us develop a better algorithm. Keep in mind that we make recommendations to be interesting and entertaining, so a certain amount of weirdness is acceptable if it also produces something inspired.

So far, only about 2,200 authors have been calculated. You can see a list of the authors here, with your authors shown in bold.

Labels: authors, recommendations

Thursday, August 25th, 2011

New tag-based recommendations algorithm

Short version. I’ve just finished up a new algorithm for calculating book recommendations based on tags. You can see them on the “Recommendations” sub-page for every work page, under a special “Tags” heading (for a limited time only!), as shown to the right. When a work doesn’t have a recommendation, it will make it. (Expect to wait 2-10 seconds.) Recommendations will be propagating through the system, and, combined with the four other recommendations algorithms we use, going into your personal recommendations over time.

Come discuss it on talk here.

Long version. Recommendation algorithms are always tricky, but doing it based on tags alone is particularly difficult. Do you consider total overlap? Overlaps by tag? How do you rate a book missing an important tag? How do you factor up meaty and meaningful tags, and downgrade meaningless, over obvious or ephemeral ones? LibraryThing has long had tag-based work-to-work recommendations, but they were of uneven quality. I haven’t been making new ones for a while now, and letting them die out of the system; all the old tag-based recommendations have been removed.

The new algorithm approaches the issue afresh, looking at all of a work’s tags, and taking into account various factors that can trip it up. It thinks about factors like tag salience on a work and generally, degree of agreement between tags and low-value tags. It also attends to similar levels of work popularity. After deciding on the basic algorithm, I toyed with various “knobs” for days, myself and with Jeremy, trying to get the best results for a set of sample and problem works. In my experience, you can’t do an algorithm like this without a sense of appropriateness, taste and proportion, and (I hope) that this is one reason LibraryThing recommendations are generally so good.

Once a tag-recommendation is generated, it takes a while for it to be incorporated into the “Combo recommendations” above. “Combo recommendations” incorporate tag recommendations to greater or lesser degrees, depending on its assessment of their quality and contribution. Your personal recommendations are based mostly on these “Combo recommendations.”

The tag recommendations are going to take a while to build for all works that need them. After that, we plan to do some sort of “Pepsi Challenge” test. We think LibraryThing recommendations are as good as any out there, and are eager to prove it.

Some examples. Tag recommendations work absolutely best on non-fiction titles about something very simple and clear cut:

Works really “about” two or more things are harder, as are books with a specific point of view, which might seem separate to some extent from the tags on it. Some examples:

  • PHP and MySQL for dummies by Janet Valade — A successful example, where most of the books deal with both PHP and MySQL, with an orientation toward “entry level” programmers (eg., the “Visual Quickstart” books)
  • Born Fundamentalist, Born Again Catholic by David B. Currie — A successful example that mostly recommends other Protestant-to-Catholic conversion stories (rather than Catholic-to-Protestant ones)
  • Goldwater by Barry Goldwater — A mixed but mostly decent result, surfacing some Goldwater-specific material, but also showing other biographies, especially from the period and involving senators. Ideally it wouldn’t have quite so many contender-bios, as I’m not sure the potential Goldwater reader is eager to dig into Edward’s bio.
  • Freakonomics by Steven D. Levitt and Stephen J. Dubner — Decent list, starting with a short-shelf of popular economics-in-life books, followed by popular introductions to economics and economics-oriented thinky-think books

Fiction, especially non-“genre” fiction, is the big problem. Literary fiction isn’t “about” what it’s about in quite the same way that non-fiction is. Creating some separation between adult and youth titles is also hard. But we’ve made significant progress.

  • Watership Down by Richard Adams — A decent list of animal-centered chapter books, with classics like Redwall and The Wind in the Willows and no The Runaway Bunny or Knuffle Bunny.
  • The Book of Three by Lloyd Alexander — A decent list of magical fantasy books, angled to youth and toward Welsh mythology.
  • The Hobbit by J. R. R. Tolkien — A good list, but obviously centered quite strongly on Tolkien. Works with significant secondary literatures (cf., Harry Potter, Narnia, etc.) tend to be dominated by guides, atlases and so forth.
  • Little Women by Louisa May Alcott — Starts out well, putting March second, but then goes off the rails somewhat after a dozen. The Mother-Daughter Book Club wins because it apparently takes place in Concord, Massachusetts and involves mothers and daughters. The Secret Life of Bees is also about mothers and daughters, coming of age, sisterhood, and was made into a movie. Meh.
  • The Book Thief by Markus Zusak — Decent list, with other books, mostly fiction, about children during the Holocaust.
  • The Shack by William P. Young — I haven’t read it, but I think have a sense of it. There are some winners here, like The Christmas List, Redeeming Love. The Year of Fog and The Deep End of The Ocean cover some of the same issues from a non-religious standpoint, and Where Is God When It Hurts covers them from a non-fiction, evangelical perspective, which might or might not be wanted. But C. S. Lewis and Paul Bunyan(!) aren’t winners, being rather different sorts of fictions. Tim LaHaye is winning almost solely on being “christian fiction,” and James Redfield for “religious fiction,” “inspirational,” etc.
  • Patient Zero by Jonathan Maberry — Zombies? We got zombies, focusing on plague-based zombie terror (no Pride and Prejudice and Zombies). Straight-up plauge titles, like The White Plague are also included. Preston’s The Demon in the Freezer probably shouldn’t be there, but it’s a great book.
  • Earth Abides by George R. Stewart — More apocalypse, this time with few zombies. The Brief History of the Dead is rather different, though I’m not sure how LibraryThing could know that. At least it doesn’t attempt to recommend other boring books, like Earth Abides.
  • Blueberry Girl by Neil Gaiman — Some good stuff. Sound lousy. The tags are dragging the recommendations all around—children, picture book, Gaiman, mothers and daughters, charles vess. Twinkle, Twinkle, Little Star is winning on “poetry,” “rhymes” and child-associated tags.
  • Love in the asylum : a novel by Lisa Carey — My wife’s book. A decent-list of insane-asylum fiction, with some memoirs.
  • Twilight by Stephenie Meyer — Largely okay, so far as I can tell, although with less secondary literature than I would have guessed.
  • Illyria by Elizabeth Hand is winning on tags like “forbidden love,” “teen romance,” “teen lit,” “romantic” and “contemporary fantasy.” I have no idea if the books are similar.

Come discuss it on talk here.

Labels: recommendations, tagging, tags

Wednesday, June 22nd, 2011

New recommendations: What should you borrow?

Member sturlington recently proposed the following:

“You know what would be a neat feature? If I could get a list of books from a friend’s library that are recommended for me based on my library. For borrowing/mooching purposes. For instance, my husband is putting in his books right now, and it would be cool to quickly get a list of his books that I might enjoy.”

I thought this an excellent idea, and easy to implement, so I did it. It’s a great way to scout out your friends’ libraries.

If you’re signed in, you can see the feature on every other (public) member’s profile page. Or you can check out what books I should borrow from Jeremy (a lot), and vice versa (a few).


Come talk about it on Talk.

Labels: recommendations

Sunday, June 12th, 2011

Recommended for you

Borrowing a feature from Netflix, I’ve introduced a “Recommended for you” section on work page. It appears whenever a book appears on the “LibraryThing Recommendations” list of any of your books.

As with other such lists, it uses the standard collection colors. It only lists books which you have marked to “include in recommendations.”

Come talk about it here.

Labels: recommendations

Monday, March 1st, 2010

Beta: “Read Alike” recommendations

I’ve pushed through a beta version of a new recommendation engine.

The “Read Alikes” recommendations supplement our existing automatic and member recommendations. “Read Alikes” are based directly on the members who have your books—the people who “read alike” you, or whatever.

So far, opinion is divided. Some members love it, and are getting great recommendations. Others report a parade of things they already know about. Is it quite consciously, however, a beta feature. It may be improved, or it may go away. Most likely, it will go away and be replaced by a better overall algorithm, with better tools for managing your recommendations.

Labels: new features, recommendations

Wednesday, December 16th, 2009

Recommending books—and series

It’s frequent for recommendations to include multiple books from the same series. So I’ve changed how recommendations display, to show these and get out of the way for other recommendations.

I also added recommendations to all Common Knowledge pages, so you can see aggregate recommendations for every series, award, place, character and so forth. For series, the results can be very good. Others can be strange, but are often quite cool.

This is another example of our continual effort to “unlearn” ecommerce, especially Amazon, conventions. Real people recommend series to each other—not to mention authors, genres, etc. Stores recommend discrete objects, because that’s what they sell. LibraryThing, which strives be interesting and useful, not to sell things, can transcend these store limitations. We just need to realize they’re there.

Come talk about it here: http://www.librarything.com/topic/79301

Labels: now with series, recommendations

Friday, January 9th, 2009

Will you like it?

I added something I’ve been working off-and-on for about a year*: “Will you like it?” Here’s an example, correctly predicting that I will like Nabokov’s King, Queen, Knave:

You’ll find the section on work pages.

Because it requires a lot of processing, you have to click to get the result. Here it is, correctly predicting that I would not enjoy a popular book about Knitting:

Each assessment has a “certainty” score (eg., “high,” “low,” etc.) based largely on how popular the book is. You can see the raw scores by hovering over the downward arrow.

How good is it? Meh. It’sokay.

This is a devilishly hard algorithm to get right. I have some ideas for improvement, but it’s fundamentally a lark and a conversation piece at present, so I don’t want to waste too much time on it.

How it works. In case you’re interested, it works completely apart from our book-to-book recommendation system, or the system that aggregates those recommendations into member-specific lists of 1,000 recommended books. Instead, “Will you like it?” works directly from the data, examining the users who have a book and how their books relate to yours.

As such, it isn’t very good at sussing out where your tastes differ from those of people who share your books. For example, my large collection of books on Greek history match me up with people who enjoy other ancient history, but I am not that interested in early Republican Rome, no matter what the algorithm thinks.

What’s interesting? I’m not going to claim it’s perfect, but it’s interesting that, to my knowledge, nobody’s every tried this before.

I think this is yet another case of Amazon limiting the horizons of what people imagine online, particularly in the online book world. Amazon pioneered book-to-book and user-to-book reviews. The work was groundbreaking but it was also routed in commercial success. User-to-book recommendations drive customers to books they’ll like and book-to-book recommendations help them find the perfect book, as well as increase the number of items in each order. Giving people honest assessments of whether they’ll like a book is murkier. Does Amazon want to tell a customer they won’t enjoy something? And what if they’re wrong?

Meanwhile, LibraryThing succeeds by being fun and interesting, not by selling books. It gives us a rare freedom to invent features that don’t sell books, like our Unsuggester—what books will you hate?—and now this.

I started a topic to discuss it.


*Don’t worry. This didn’t distract. I just pushed two combination/separation bug fixes, and Chris and I are hard at work on the catalog, in preparation for some larger changes (ETA: one week?).

Labels: amazon, new features, recommendations

Monday, May 26th, 2008

LibraryThing recommendations!



LibraryThing Recommendations—called “the best feature on the site” by one user—are back and much better than before.

You can find recommendations at the top of your profile page. Or check out mine.

The new recommendations include:

  • A large number of primary recommendations for ever member—usually 1,000—based on a single comprehensive algorithm.
  • Individual recommendation lists for each member’s tags.
  • Filtering of recommendations by popular LibraryThing tags.
  • Individual lists of other members’ recommendations (member recommendations were added two weeks ago)
  • Up to 500 so-bad-they’re-good recommendations, building off the LibraryThing Unsuggester, and called “Your Unsuggester.”* We hope “What I shouldn’t read” has some meme legs.
  • A “why” feature for each recommendation, laying out what the recommendation was based on.
  • A pony.**

I let the recommendations themselves out early—see the original talk post, with over 140 messages!—and members had mostly positive reactions. Those who don’t like them can perhaps be molified by the greater number and ways to filter and angle the recommendations.

Recommendations now change daily—faster if you are below 200 books and keep adding them. The system keeps track of all recommendations and when you received them. In the near future I plan to provide personalized recommendation emails based on new recommendations.

I’ve created a new Talk thread to discuss the changes, and suggest changes. My thanks to those who participated in the initial thread, influencing development in a number of important ways.


*If Thomas Jefferson is in Hell, I am confident the Devil is torturing him with books from Jefferson’s Unsuggester List—heavy on the chick- and tween-lit!
*With apologies to Last.fm.

Labels: new feature, new features, recommendations

Monday, May 19th, 2008

Member book recommendations

I’ve just added a small new feature, Member recommendations. You can check it out under “Recommendations” here or here.*

Basically, you can now add your own recommendations to LibraryThing’s six (!) algorithmic recommendations. If you want, you can also leave a short explanation of your choice.

I’m throwing this one out pretty raw.** It’s available from the primary page of a work, and from its recommendation page, and on a single Member Recommendations page.

To be done:

  • A way to see all the recommendations you’ve given
  • A way to see all the recommendations others have applied to your books
  • Recommendation flagging
  • Up/down voting on recommendations?

Come talk about the feature and where it could go on Talk here.


*I hope to link to some better examples soon, one members start adding them. I find fiction recommendations very hard, so most of my recommendations so far have been off ancient history, which makes the feature seem much less interesting than it is!
**I’ve had this on ice for a while, while dealing with tags and scaling issues. I don’t think I’m going to be making major changes until Chris comes back from paternity leave later this week or next.

Labels: new feature, new features, recommendations