Music Suggestion System

CS 250 Final Project

Matt Jones, Travis Thomas, Xingbang Liu

Motivation

Why Music Suggestion System?

• Build off of previous labs
• Spotify & Pandora

Algorithms

• Memory-based

$$cos(A) = \frac{\vec a \cdot \vec b} {|\vec a| \times |\vec b|}$$

• Model-based

$$R_{user,song}= Q_{user,factor} P^T_{factor,song}$$

Background/Overview

Repository structure overview

Like List

• liked.csv
• BufferedReader
• Prompt out titles

• String[] Songs = line.split(cvsSplitBy);

• Sort by int(bubble sort) or String(exchange sort)

Explore

• Store our Cloud list (music.csv) into a 2D ArrayList
• song<>(String) in exploreSongs<>(String)
• randomly pick 5 songs
• if user likes, then update "y" (in music.csv) by using BufferedWriter
• add line to liked.csv by using FileWriter

Music Suggestion

• Not done yet
• Melody score (1-10 representing speed)
• Calculate mean score from liked.csv
• Linear search in music.csv
• Randomly select 2 songs

Math Evaluation

• By running

  System.nanoTime()

  four times, I got:

• in nano seconds	bubble sort	exchange sort
  1.	186,525	106,856
  2.	156,695	111,009
  3.	156,318	110,631
  4.	239,386	110,254
  mean	184,731	109,688

• $$O(n^{2})$$

Conclusion

• Making good progress
• inheritance
• Biggest Challenge - Overwriting Y in the music.csv

2d arrayList, Test.csv

Do you have any questions?

. . .