Introduction to Text Classification in R with quanteda

提供:
Coursera Project Network
このガイド付きプロジェクトでは、次のことを行います。

Import text documents, reshape texts from documents to paragraphs, and turn your texts into a machine readable format.

Classify presidential concession speeches by political party using a Naive Bayes algorithm and assess the accuracy of the predictions.   

Clock2 hours
Beginner初級
Cloudダウンロード不要
Video分割画面ビデオ
Comment Dots英語
Laptopデスクトップのみ

In this guided project you will learn how to import textual data stored in raw text files into R, turn these files into a corpus (a collection of textual documents), reshape them into paragraphs from documents and tokenize the text all using the R software package quanteda. You will then learn how to classify the texts using the Naive Bayes algorithm. This guided project is for beginners interested in quantitative text analysis in R. It assumes no knowledge of textual analysis and focuses on exploring textual data (US Presidential Concession Speeches). Users should have a basic understanding of the statistical programming language R.

あなたが開発するスキル

  • Ordered Pair
  • Text Analysis
  • Algorithms
  • Statistical Programming Languages
  • Computer Programming

ステップバイステップで学習します

ワークエリアを使用した分割画面で再生するビデオでは、講師がこれらの手順を説明します。

  1. Load text documents into R studio, convert a number of text documents into a corpus, and extract data from text document file names and add them to a new column in a dataframe. 

  2. Reshape the dataset into paragraphs from documents and check for balance in your labels. 

  3. Split up a text document corpus into tokens, or individual words and punctuations. Then clean the data by removing specific words and spellings.

  4. Create a Document Feature Matrix, divide it into train and test sets and run a Naive Bayes model. Then examine the model’s prediction accuracy and learn about accuracy, precision, and recall.   

  5. Run Naive Bayes models for a second and third time. Then examine the models' predictions and compare the model outputs with results from the previous task.

ガイド付きプロジェクトの仕組み

ワークスペースは、ブラウザに完全にロードされたクラウドデスクトップですので、ダウンロードは不要です

分割画面のビデオで、講師が手順ごとにガイドします

よくある質問

よくある質問

さらに質問がある場合は、受講者ヘルプセンターにアクセスしてください。