【IT】Introduce User Local Inc. emotion recognition AI service / 「ユーザーローカル感情認識AI」サービス

in utopian-io •  6 years ago 

freestocks-org-470416-unsplash.jpg

1. Introduction

User Local Co., Ltd., which is strong in the fields of data processing and machine learning Corporation, released a Web service "User Emotion Recognition AI" that determines emotions from texts entered on December 17.

データ処理や機械学習の分野で強みのある株式会社ユーザーローカルは12月17日、入力したテキストから感情を判定するWebサービス「ユーザーローカル感情認識AI」を公開しました。

1.png

2. Overview

In this service, when entering text into a form, you can quantify the five emotions of "joy" "liking" "sadness" "fear" "anger" contained in it.

無料で利用でき、このサービスではフォームにテキストを入力すると、そこに含まれる「喜び」「好き」「悲しみ」「恐れ」「怒り」の5項目の感情を数値化することができます。

Let's try to see how joyful I am with steepian my articles.

自分の記事にどれほど喜びがあふれているか、試してみたいと思います!

3. Actual Tutorial Content

3.1.Determine emotion of input sentence / 入力された文章の感情を判定

First, let's try to judge emotions of sent sentences through "User Local Emotion Recognition AI" service. I will explore the feelings of "French restaurant standing quietly behind the alley marugame" by @tamito0201 wrote yesterday.

まず入力された文章の感情を判定してみます。私、@tamito0201が書いた「【フレンチ】東京・神楽坂の「マルガメ」。路地裏にひっそりと佇む隠れ家で今宵も一杯を楽しむ!」の感情を探ってみます。

I entered the following sentence because it is within 300 words.

300字以内ですので以下の文章を入力しました。

(English)
Wine and dinner quietly standing in the back alley of Kagurazaka Oguri-dori "marugame (marugame)". It is a restaurant like a hideout where you can enjoy a small casual French with 6 counters and 2 tables. We prepare dinner instead of the wine to match the wine and enjoy the cuisine while tilting the glass to enjoy the evening meal of Mr. Margo every day at the counter Everyone male and female do not care about Margua fan who enjoys cooking. I ordered the course in advance on this day. I will have dinner slowly. Taste is over matsutake mushrooms, cucumber, mushrooms and Hanabiratake. A campus of yam's sprout bud and Hanabiratake is spreading on the plate in valuable natural mushroom incense mushrooms growing only in a few mountains.
(Japanese)
神楽坂小栗通りの裏路地にひっそりと佇むワインと夜ご飯「マルガメ(marugame)」。カウンター6席、テーブル2席の小さなカジュアルフレンチが楽しめる隠れ家のようなお店です。ワインに合う月替わりの夜ご飯を用意していてカウンターにも連日そのマルガメさんの夜ご飯を楽しもうと一人でもグラスを傾けながらお料理を楽しむ老若男女問わないマルガメファンが押し寄せています。この日は事前にコースを注文。ゆっくりと夜ご飯を頂きます。むかご、香茸、ハナビラタケをそえて美味しさは松茸以上。ごく一部の山にしか生えない貴重な天然キノコ香茸に山芋の球芽むかご、ハナビラタケのキャンパスがお皿の上に広がっています。

Results with sentences with strong emotions of your choice.

好きの感情が強い文章との結果がでました。

2.png

Today is Christmas Eve no matter what. I want to judge how much everyone in the world is full of love.

今日はなんといってもクリスマスイブ。世の中のみんながどれくらい愛に溢れているか判定したいと思います。

2.jpg

  • Miss order to Santa who regret delivered a different one / 違うものが届いて後悔したサンタさんへの発注ミス
(Aichi 27-year-old man in Japan)
I wanted a radio control of a tank from which BB bullets appeared, but I only asked for "radio control of a tank", and actually I came that type that BB bullets did not come out. I regretted that I should have told it properly
(愛知・27歳男性)
BB弾が出る戦車のラジコンが欲しかったのですが、『戦車のラジコン』とだけお願いしてしまい、実際に来たのはBB弾が出ないタイプでした。ちゃんと伝えておけば良かったと後悔しました。

3.png

I distinguished AI as a fearful sentence. This sentence is a sentence closer to sadness than fear, but since there are characters such as tanks and bullets, it may be that AI has determined "fear".

恐れいる文章とAIが判別しました。恐れているより悲しみに近い文章ですが、戦車や弾という文字がありますので"恐れ"とAIが判別したのかもしれません。

I also use text mining tools.

続いてテキストマイニングツールも使っています。

3.2 text mining tools / テキストマイニングツール

3.2.1 Word cloud / ワードクラウド

I would like to text mining the following articles I wrote.

[Contest] @ steemit-jp Contest # 26 "Christmas! What gift would you like to ask Santa?"

自分が書いた以下の記事をテキストマイニングしてみたいと思います。

【コンテスト】@steemit-jpコンテスト#26 「クリスマス!サンタさんにお願いしたいプレゼントは何?」

4.png

I see. It is a sentence to which the desperation whom desire swirled so much is transmitted.

なるほど。とても欲望が渦巻いた必死さが伝わってくる文章で御座います

5.png

3.2.2 Word appearance frequency / 単語出現頻度

Frequency of word occurrence. "Present" 6 times, "Christmas" 4 times, "Wish" 4 times has appeared.

単語の出現頻度です。「プレゼント」が6回、「クリスマス」が4回、「お願い」も4回登場しています。

6.png

どうでしょう、どなたか僕にクリスマスプレゼントを恵んでくれてもよいのではないでしょうか。

How is it that someone may endorse Christmas gifts for me?

3.2.3 Co-occurrence keyword/ 共起キーワード

It is a figure where words appearing in sentences have similar appearance patterns connected by lines.

文章中に出現する単語の出現パターンが似たものを線で結んた図です。

7.png

It is associated with my wish that "gifts" and "please" and "go to" to "Kyoto" are associated!

"プレゼント"と"お願い"、"京都"へ"行く"が関連付けられているとこが僕の願いと一致していますね!

3.2.4 2D map / 二次元マップ

Words with similar appearance tendencies in sentences are closer, and words that are not similar are closer together

文章中での出現傾向が似た単語ほど近く、似ていない単語ほど遠く配置されています。

8.png

I am close to the keyword "London hit" and "travel won tickets", "I want it". Looking at this figure is nice because it falls into the illusion that somehow got a present.

ロンドンパスや旅行券が"当たる"、"欲しい"というキーワードの近くにいます。この図を眺めているとなんとなくプレゼントに当たったような錯覚に陥りるからナイスです。

There are other "dependency analysis" and "hierarchical clustering" services, but user registration is required.

他に「係り受け解析」、「階層的クラスタリング」のサービスもありますが、ユーザー登録が必要です。

3.3 Automatic summary tool / 自動要約ツール

It is a tool that literally summarizes sentences. I would like to summarize the same sentences as before text mining.

文字通り文章を要約してくれるツールです。さきほどのテキストマイニングと同じ文章を要約してみたいと思います。

9.png

It is the result of 3 row summarization./ 3行要約の結果です。

10.png

I did a bear with a gift box behind. I had you take a two-shot with a bear "Detective Pikachu "London bus charter tour in Japan".


こちらはプレゼントボックスを背景にクマさんを致しました。 クマさんとのツーショットを撮ってもらいました "「名探偵ピカチュウ」ロンドンバス貸切ツアー in Japan "。

In summary it seems that it is going to be "" Detective Pikachu "London bus charter tour in Japan." Oh, I went to a London bus charter tour again!

要約すると"「名探偵ピカチュウ」ロンドンバス貸切ツアー in Japan "に行ったことになるようです。そうか、もうロンドンバス貸切ツアーに行ったんだ!

I was able to fully enjoy Christmas Eve. hahaha ...

いやぁクリスマスイブをとことん満喫できましたよ。え?

4. Conclusion

A Web service "user local emotion recognition AI" for judging feelings from inputted text. It might be interesting to try Steemian's emotions as well as writing sentences written in blogs and SNS.

如何でしょうか。入力したテキストから感情を判定するWebサービス「ユーザーローカル感情認識AI」。皆様もブログやSNSで書いた文章がどのような感情で書かれているか試してみるのも面白いかもれません。

It seems interesting to try it with various news articles etc.

色んなニュース記事などで試してみるのも面白そうですね。

line2.gif

written by tamito0201

Block chain & web engineer.

A free engineer engaged in the development of business application in the fields of web applications and finance, securities, communication, etc. After that, I experienced building virtual currency platform, I fascinated the development of DAO and DApps using block chain technology, and keep on hacking every day.

Upvote, Follow, Resteem of everyone will encourage renewal, so thank you for your support.

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
Sort Order:  

Thank you for your contribution @tamito0201.
After analyzing your tutorial we suggest the following:

  • While this can come in handy, basic on screen instructions are normally not the best content we look for under utopian tutorials.

  • The tutorial gets cleaner if you first write in English and then at the end write the tutorial again in your language.

  • Please read this link.


Need help? Write a ticket on https://support.utopian.io/.
Chat with us on Discord.
[utopian-moderator]

Thank you for pointing out carefully.

I would like to read the link carefully. I would like to write good quality articles so that I can continue to contribute to Utopian-io.

Thank you for your support.

Thank you for your review, @portugalcoin! Keep up the good work!

Thank you for everyone, reviews and comments.

I will try to apply good quality articles!

Nice picture!!!

Merry Christmas! I will write helpful information in the future!

Congratulations @tamito0201! You have completed the following achievement on the Steem blockchain and have been rewarded with new badge(s) :

You received more than 4000 upvotes. Your next target is to reach 5000 upvotes.

Click here to view your Board
If you no longer want to receive notifications, reply to this comment with the word STOP

Do not miss the last post from @steemitboard:

Christmas Challenge - Send a gift to to your friends

Support SteemitBoard's project! Vote for its witness and get one more award!

Congratulations @tamito0201! You have completed the following achievement on the Steem blockchain and have been rewarded with new badge(s) :

You published a post every day of the week

Click here to view your Board
If you no longer want to receive notifications, reply to this comment with the word STOP

To support your work, I also upvoted your post!

Do not miss the last post from @steemitboard:

Christmas Challenge - Send a gift to to your friends

Support SteemitBoard's project! Vote for its witness and get one more award!