Patient’s Condition Classification Using Drug Reviews

Table of Contents:

1. Description

2. Dataset

3. EDA - Exploratory data analysis

4. Data Preprocessing

4. Model

5. Evaluation

1. Description

We build a system that can identify patient's condition by the help of both Natural Language Processing(NLP) and Machine Learning(ML) in classifying patient to reduce the efforts and time expanded by the doctors and evaluate the type of patient at an early stage.

2. Dataset

The datset is collected from UCI.

3. EDA - Exploratory data analysis

  * Statisticaly analysis data

4. Data Preprocessing

1. Sopt Word
2. Lemmatization
3. Split the dataset
   • Split the dataset with 80% of training set and 20% of test set.
4. Creating features and Target Variable

4. Model

1. TF-IDF - TF-IDF is a very popular feature extraction technique. Text needs to converted into vector or matrix before fed them to the Machine Learning model.
2. Bag of Words
3. Naive Bayes
4. Passive Aggressive Classifier

5. Evaluation

1. Confusion Matrix
2. Accuracy