We will explore the forecasted values of based on n ranging from 1 to 23 and k

We will explore the forecasted values of based on n

This preview shows page 2 - 5 out of 5 pages.

Where n+1 is the dimensionality of the model.We will explore the forecasted values of based on “n” ranging from 1 to 23 and “k”ranging from 1 to 10, where n determines the dimension of the model and k is the number of nearest neighbors. Working:To predict number of google searches, we have used Excel to build KNN Models for the data.By exploring models with n=1,3,….,23 we have following outputs:
Image of page 2
Image of page 3
Image of page 4
As seen in the above table the closest prediction for the number of google searches for “pizza” on 9th January 2020 at 8:00 PM is 41.75 and 42.25 where (n, k) are (2,4), (2,8), (9,8), (10,8), (11,8) …., (23,8). Findings: 1. The most accurate prediction of google searches for the term “Pizza” on 9th January 2020 at 8:00 PM is 41.75 and 42.25 at (n,k) = (2,4), (2,8), (9,8), (10,8), …., (23,8) where the actual value is 42, KNN model is 99.5% accurate for this value. References: 1. Introduction to k-Nearest Neighbors: A powerful Machine Learning Algorithm (with implementation in Python & R) , Tavish Shrivastava, 2018 (Retrieved from: - neighbors-algorithm-62214cea29c7) 2. A Quick Introduction to K-Nearest Neighbors Algorithm, Adi Bronshtein, 2017 (Retrieved from: - algorithm-clustering/)
Image of page 5

You've reached the end of your free preview.

Want to read all 5 pages?

  • Winter '20

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture