NLP End Term Exam

December 14, 2019
The below Experiment is Based on two way Neural Network Model with Tune parameters also
Answers: 1) **Accuracy** for the **Two Class Neural Network** with default parameters is **0.615** 2) **Accuracy** for the **Two Class Neural Network by** **Tuning Hyper Parameters** is **0.637** 3) **Accuracy** for the **Two Class Neural Network** by Tuning Hyper Parameters with **two Hidden Layers** : a. When **auto** functions for hidden layers: **0.637** b. When **Manual** functions for hidden layers i.e **[50]** and **[60]**: **0.643** 4) I have created the model for recommendation also. a. **Variables** we used for Recommender system are: i. reviewerId ii. asin iii. overall The Data structured in the specific format as: **(reviewerID, asin, overall)**. So we get the recommendation of the rating for each of the reviewer through its ID. NDCG we get is 0.947083. So on the basis of reviewerId we will get the overall rating of accuracy 94.70% 5) The wordcloud has been created from the corpus with 100 max words. a. We can say that the **use** and **work** word is used maximum from the corpus. As most of the reviews are based on the uses of the lawn and garden ![Word Cloud][1] [1]: http://file:///C:/Users/HUDA/Downloads/Desktop/Wordcloud.png