post your websites

mnl wrote

Trying to build an Anime recommender system, based on the MyAnimeList kaggle dataset with the plan to eventually host it on a website. So far, I've spent the day hacking together a short script that at already kinda works, although I haven't finetuned anything yet.

I created "recommendations" for myself based on my profile in the dataset, but as the dataset is already more than a year old and as I am notoriously bad at keeping my animelist up to date it's mostly stuff I've already watched.

The top 50 "recommendations" for my (outdated) profile:

Cowboy Bebop, Death Note, Fullmetal Alchemist: Brotherhood, Toki wo Kakeru Shoujo, Howl no Ugoku Shiro, Samurai Champloo, Byousoku 5 Centimeter, Ghost in the Shell, Bakemonogatari, Akira, Tonari no Totoro, Evangelion: 1.0 You Are (Not) Alone, Higurashi no Naku Koro ni, Summer Wars, Mushishi, Hotaru no Haka, Paprika, Evangelion: 2.0 You Can (Not) Advance, Clannad, Serial Experiments Lain, Clannad: After Story, Trigun, Kaze no Tani no Nausicaä, Fate/Zero, Toradora!, K-On!, Angel Beats!, Higashi no Eden, Ergo Proxy, Cowboy Bebop: Tengoku no Tobira, Perfect Blue, Lucky☆Star, Great Teacher Onizuka, Soul Eater, Shingeki no Kyojin, Tenkuu no Shiro Laputa, Ano Hi Mita Hana no Namae wo Bokutachi wa Mada Shiranai., Fate/Zero 2nd Season, Ghost in the Shell: Stand Alone Complex, Kara no Kyoukai 2: Satsujin Kousatsu (Zen), Hellsing Ultimate, Higurashi no Naku Koro ni Kai, Monster, Kara no Kyoukai 5: Mujun Rasen. Majo no Takkyuubin, Kara no Kyoukai 3: Tsuukaku Zanryuu, Psycho-Pass, Suzumiya Haruhi no Yuuutsu (2009), Sayonara Zetsubou Sensei, Dragon Ball Z

Certainly some good recomendations. I watched most of the stuff and enjoyed pretty much everything,  so I guess that shows that the system already works quite well? Another possibility is that it's just recommending me stuff everybody likes.

Next up:

  • webscrape the up-to-date animelist of users I want to build recommendations for, so that I  can remove animes that are already on their list from the recommendations
  • maybe I should scrape together a whole new up-to-date dataset while I am at it
  • atm I am using matrix factorization for predicting scores users would give animes they haven't watched yet. Of course that only works for users that are already part of the dataset, so I need to think of a way to predict scores of users not part of the dataset

This is some high-level stuff... I love the concept, though! And it sounds like you're over halfway there. I've been notoriously bad at keeping my programming skills sharp lately, but if you'd like, I'd love to see the site and beta-test someday. ^^

Backlash wrote
mnl wrote

Trying to build an Anime recommender system, based on the MyAnimeList kaggle dataset with the plan to eventually host it on a website. So far, I've spent the day hacking together a short script that at already kinda works, although I haven't finetuned anything yet.

I created "recommendations" for myself based on my profile in the dataset, but as the dataset is already more than a year old and as I am notoriously bad at keeping my animelist up to date it's mostly stuff I've already watched.

The top 50 "recommendations" for my (outdated) profile:

Cowboy Bebop, Death Note, Fullmetal Alchemist: Brotherhood, Toki wo Kakeru Shoujo, Howl no Ugoku Shiro, Samurai Champloo, Byousoku 5 Centimeter, Ghost in the Shell, Bakemonogatari, Akira, Tonari no Totoro, Evangelion: 1.0 You Are (Not) Alone, Higurashi no Naku Koro ni, Summer Wars, Mushishi, Hotaru no Haka, Paprika, Evangelion: 2.0 You Can (Not) Advance, Clannad, Serial Experiments Lain, Clannad: After Story, Trigun, Kaze no Tani no Nausicaä, Fate/Zero, Toradora!, K-On!, Angel Beats!, Higashi no Eden, Ergo Proxy, Cowboy Bebop: Tengoku no Tobira, Perfect Blue, Lucky☆Star, Great Teacher Onizuka, Soul Eater, Shingeki no Kyojin, Tenkuu no Shiro Laputa, Ano Hi Mita Hana no Namae wo Bokutachi wa Mada Shiranai., Fate/Zero 2nd Season, Ghost in the Shell: Stand Alone Complex, Kara no Kyoukai 2: Satsujin Kousatsu (Zen), Hellsing Ultimate, Higurashi no Naku Koro ni Kai, Monster, Kara no Kyoukai 5: Mujun Rasen. Majo no Takkyuubin, Kara no Kyoukai 3: Tsuukaku Zanryuu, Psycho-Pass, Suzumiya Haruhi no Yuuutsu (2009), Sayonara Zetsubou Sensei, Dragon Ball Z

Certainly some good recomendations. I watched most of the stuff and enjoyed pretty much everything,  so I guess that shows that the system already works quite well? Another possibility is that it's just recommending me stuff everybody likes.

Next up:

  • webscrape the up-to-date animelist of users I want to build recommendations for, so that I  can remove animes that are already on their list from the recommendations
  • maybe I should scrape together a whole new up-to-date dataset while I am at it
  • atm I am using matrix factorization for predicting scores users would give animes they haven't watched yet. Of course that only works for users that are already part of the dataset, so I need to think of a way to predict scores of users not part of the dataset

This is some high-level stuff... I love the concept, though! And it sounds like you're over halfway there. I've been notoriously bad at keeping my programming skills sharp lately, but if you'd like, I'd love to see the site and beta-test someday. ^^

Thanks for the offer! I'll definitely come back to that and will also keep you guys updated on my progress. I'm coming from a statistics background, so the biggest challenge will be the webdev/database/deployment stuff and not the actual machine learning. I reckon it will take quite a while... ^^

10 days later

Vastly improved my model and moved training to an university server (took 2.5h on 40 CPUs).

My personal predictions:

Top 10 Anime Predictions
	(10.9813) Boku no Pico
	(10.6180) Shitcom
	(10.4071) Pico to Chico
	(10.3110) Pico x CoCo x Chico
	(9.9006) Pico: My Little Summer Story
	(9.8840) Pussycat
	(9.8479) Otome Dori
	(9.7576) No Littering
	(9.7455) Bloody Date
	(9.7291) Enzai

Bottom 10 Anime Predictions
	(5.9809) Hayate no Gotoku!
	(5.9607) Clannad Movie
	(5.9427) Suzumiya Haruhi no Yuuutsu (2009)
	(5.9360) Ookami to Koushinryou II
	(5.9072) Kore wa Zombie Desu ka?
	(5.8592) Natsume Yuujinchou
	(5.8119) Tsubasa Chronicle 2nd Season
	(5.6986) Cossette no Shouzou
	(5.5498) Chobits
	(5.4300) xxxHOLiC

Well... The model certainly improved but I must have gotten a sign wrong somewhere... facepalm

For those interested, here is a list of features with the top and bottom ranking anime in each feature:

Feature 1:
	max : (+4.0) Suzumiya Haruhi no Yuuutsu
	min : (-6.0) Naruto: Shippuuden
Feature 2:
	max : (+5.0) .hack//Sign
	min : (-3.1) Tsubasa Chronicle 2nd Season
Feature 3:
	max : (+4.7) Sword Art Online
	min : (-3.3) Higashi no Eden Movie II: Paradise Lost
Feature 4:
	max : (+7.0) Pingu in the City
	min : (-4.1) Lucky☆Star
Feature 5:
	max : (+4.5) Bobobo-bo Bo-bobo
	min : (-4.4) Aki-Sora
Feature 6:
	max : (+5.9) Ore no Imouto ga Konnani Kawaii Wake ga Nai. Specials
	min : (-5.2) School Days
Feature 7:
	max : (+3.6) Boku no Pico
	min : (-7.6) School Days
Feature 8:
	max : (+3.6) Vampire Knight: Guilty
	min : (-3.9) Hachimitsu to Clover
Feature 9:
	max : (+5.4) Dragon Ball GT
	min : (-4.1) Aku no Hana
Feature 10:
	max : (+6.9) One Piece
	min : (-2.9) Princess Princess
Feature 11:
	max : (+4.2) School Days
	min : (-5.7) Sword Art Online II
Feature 12:
	max : (+4.6) Neon Genesis Evangelion
	min : (-4.2) Shinsekai yori
Feature 13:
	max : (+4.3) Bokusatsu Tenshi Dokuro-chan
	min : (-3.7) K-On!!
Feature 14:
	max : (+5.1) Pupa
	min : (-4.4) Tenshi Kinryouku
Feature 15:
	max : (+4.5) Poputepipikku
	min : (-5.3) Aku no Hana
Feature 16:
	max : (+5.4) Aku no Hana
	min : (-3.6) Green Green
Feature 17:
	max : (+3.4) Texhnolyze
	min : (-5.9) Suzumiya Haruhi no Yuuutsu (2009)
Feature 18:
	max : (+4.5) Diabolik Lovers
	min : (-7.2) School Days
Feature 19:
	max : (+3.7) Dragon Ball Z Movie 11: Super Senshi Gekiha!! Katsu no wa Ore da
	min : (-5.7) Boku no Pico
Feature 20:
	max : (+3.8) Dragon Ball Super
	min : (-3.4) Pupa
Feature 21:
	max : (+4.6) Neon Genesis Evangelion: The End of Evangelion
	min : (-4.5) Ore no Imouto ga Konnani Kawaii Wake ga Nai. Specials
Feature 22:
	max : (+2.5) Poputepipikku
	min : (-6.6) Berserk
Feature 23:
	max : (+6.2) Pico x CoCo x Chico
	min : (-4.5) Dragon Ball GT
Feature 24:
	max : (+4.4) Bokura ga Ita
	min : (-4.5) Angel Beats!
Feature 25:
	max : (+2.8) Yu☆Gi☆Oh! 5D's
	min : (-5.2) Aki-Sora: Yume no Naka
Feature 26:
	max : (+4.3) Charlotte
	min : (-4.1) Bible Black
Feature 27:
	max : (+4.7) Boku no Pico
	min : (-4.4) Evangelion: 3.0 You Can (Not) Redo
Feature 28:
	max : (+5.1) Bokusatsu Tenshi Dokuro-chan
	min : (-3.9) Dragon Ball Z
Feature 29:
	max : (+4.7) Sword Art Online II
	min : (-3.3) Full Moon wo Sagashite
Feature 30:
	max : (+3.4) Fate/Zero
	min : (-4.5) Inferno Cop
Feature 31:
	max : (+6.3) School Days
	min : (-4.2) Katekyo Hitman Reborn!
Feature 32:
	max : (+3.9) Dragon Ball GT
	min : (-5.4) Diabolik Lovers
Feature 33:
	max : (+5.2) Naruto: Shippuuden
	min : (-4.6) Bobobo-bo Bo-bobo
Feature 34:
	max : (+3.7) Black★Rock Shooter (TV)
	min : (-4.1) Yosuga no Sora: In Solitude, Where We Are Least Alone.
Feature 35:
	max : (+4.5) Bokusatsu Tenshi Dokuro-chan
	min : (-3.7) Byousoku 5 Centimeter
Feature 36:
	max : (+5.1) Gintama
	min : (-3.6) Ouran Koukou Host Club
Feature 37:
	max : (+3.8) Air
	min : (-5.3) Highschool of the Dead
Feature 38:
	max : (+3.8) Vampire Knight: Guilty
	min : (-3.7) Gintama
Feature 39:
	max : (+3.6) Ousama Game The Animation
	min : (-4.6) To LOVE-Ru
Feature 40:
	max : (+9.5) School Days
	min : (-3.5) Boku no Pico
mnl wrote

Well... The model certainly improved but I must have gotten a sign wrong somewhere... facepalm

You sure? Looks about right imo

4 years later

Hi there,
First of all I would like to say that LEWD is looking pretty good, I like your activity feed.

Our site Gamer.Express is a gaming forum which is mainly dedicated to discussing games, but we also have a section for anime fans.

Right now we don't have a person in charge of the anime section, so it would be great to hear your recommendations. I see that there are a lot of experienced members here so any advice would be appreciated.
😅

    Journey A lot of experienced members, sure. A lot of experienced members here? I wish ^_^'

    Alas, I don't have any recommendations for how to get an anime forum off the ground, except maybe to advertise (like you mid-key did here~). I think most people got here through ads on relevant websites, or through other forums.

        10 days later
        2 months later
        2 years later

        Been lurking here since 2020, finally have something interesting to share.

        Launching: Anihwa

        Think Twitter meets Discord for anime-like things.
        Would love to hear what you guys think, there's a lot more to come.

        Fun fact it's flarum under the hood, extensively modified.

          Excalius very sleek. I don't watch anime enough to want to talk about it all the time though. So you're saying this isn't just a wrapped website?

              COUPLED It isn't just about anime or art It really encompasses the broader topics. Anime and art has a stronger presence because of current trends and audience interest, but it's a focus not the entire site's identity.

              I have a huge roadmap for this site, just to name a few once traction builds.

              1. Marketplace for creators/artist to sell their creations online such as comics, novels etc.. and read them directly through the site.
              2. User created content to use on the site like emoji packs, profile themes for page/avatar similar to what discord offers but created by the community.
              3. Point Shop, e.g. daily login, complete task, use to purchase stuff from the marketplace.
              4. And much more :3 See Roadmap

                  Excalius Looks cool! I'm not that into art either but it looks like it would be a fun place for artists.

                      darpo It's alot more than that, granted yes artisit will for sure appreciate it. It's a forum at the end of the day there's more topics than the highlighted ones.

                      I got a lot to do.

                        5 days later

                        Mine is Blushcast: https://blushcast.com/

                        Full disclosure: I’m involved with the project. It is an 18+ library that brings original adult fiction, audio stories, guided audio and interactive stories together in one place.

                        The interactive side is the part I’m most interested in at the moment. The aim is for choices to change the conversation, pacing or ending instead of only changing one sentence before returning the reader to the same route.

                        All characters are adults and the content is fictional.

                        I would be curious which format people here actually use most: written stories, audio, or branching interactive fiction?

                        COUPLED I don't reacall having a points system... I do recall having love hearts instead of likes, but I guess you can still leave the former~