Understanding Semantic Search

Jim Morrison was dead and the Doors would never open again,  Jimmy Hendrix was playing through a Purple Haze in heaven, and the Lord bought Janis a one-way ticket, now she has colour TV no doubt…I was 16 and searching for music that would sustain me for the rest of my life.

The Semantic Search Begins

So I ask my buddy Google (that’s what we called him, I swear) if he could find me some music to listen to.  I didn’t have a lot of time or money to buy a whole bunch of albums, most of which I would only listen to once. So Google set off on his task, being the friend that he was.  He was also a little weird, he just liked to find things out.  So off he went and asked what music these people listened to:

  • my closest friends
  • their friends
  • people that looked like me,
  • people who did the same things I did

and he came up with a REALLY BIG LIST!

Ordering the list

Big Data Points

Now Google was a really good friend and he didn’t want to just give me the list he had compiled, he wanted to put it in order so the top of the list was going to be the most probable choices of me liking the music.  So he went out and asked , “Who knows the most about music.”

Google asked, “Who knows the most about music?”

He came up with a couple names of people I knew:

  • Kenny, a DJ that talked about music all day long.
  • Dee-Dee, who everyone liked and listened to music all the time.

Then Google went outside of the people I knew,  he called the radio stations and asked about the music they played, and talked to the radio DJ’s, and maybe indulged in the euphoric pleasures of the herb backstage with some bands and he got to know what the “good” music was with all these opinions.  The problem was that they all had different opinions – so who’s opinion should he give the most weight to when ordering the list.

Veracity

Google noticed that when Kenny played something new at the school dance everyone noticed and ran out to buy the album because Kenny said it was good.  Kenny’s opinion was so valued that the DJ’s in neighbouring towns would play the same music the following week and the news spread quickly when Kenny played something new, so Google thought that Kenny’s opinion was important.

Next he looked at Dee-Dee (everyone liked looking at Dee-Dee) and even though everyone listened to Dee-Dee nobody shared her opinion with their friends – Dee-Dee was just nice, and easy to look at, but she changed her mind about music everyday.  Everyone agreed she had great pictures of the bands from backstage but her opinions about the actual music where wishy-washy.

Finally Google considered the opinions of the Radio Station DJ’s and the Rock Band Players.  The latter really only liked their own music while the former, Google suspected, was being paid to play certain bands.

So Google decided that Kenny’s opinions had the greatest Veracity and therefore he lent some of that veracity to Kenny’s musical choices.

Volume & Variety

Now the radio DJ’s definitely played the most music, even more than Kenny.  So he had to consider this – they were exposed to a lot more music than Kenny so he had to give some weight to their opinions.  The other thing he noticed was that bands with a greater Volume of music got played more, just because they made more music.  This led to another discovery that some bands had great Volume but only got played on one radio station – and Kenny would only play a couple of their songs at the dances, while other bands got played on multiple radio stations and Kenny played more of their music at the dances because they had a greater Variety of music.  Google further surmised that bands that could play a greater variety of music understood the subject, music, better than those that didn’t.

Velocity

Now the list was really shaping up, but Google really wanted to fine tune it, so he thought and he thought and came up with the idea that the faster a song became popular, the better it must be.  He looked at what the radio DJ’s played and saw that a song’s popularity built steadily but slowly.  On the other hand, if Kenny played a song, locally, the song could be heard playing everywhere in just a week.  The songs that Dee-Dee liked might skyrocket to popularity quickly, but a week after, they were never heard again.

So Google ordered his list, giving those bands with the greatest Velocity in popularity a little boost upward on the list.

Putting it all together

So after deep debate Google considered the Veracity of the music and the Veracity of the proponents of the music.  At the same time he considered the Volume  of the music, the Volume it was played and shared.  He also considered the Volume that the different people shared in general and if they shared a lot about music, their opinions lent credence to the music’s Veracity and as a side effect increased the music’s Velocity in popularity.  While this was all going on he also considered the Variety of music the bands played and the Variety of people that played the music and the Variety of recommendations that each proponent made about music in general.

After this exhausting exercise where Google considered 110,000,000 data points in 0.37 seconds this is what he came up with:

Led Zeppelin – barely made it to first position.  They had volume, variety and veracity but their velocity was slowing by 1976 it only made it to #1 because all my friends loved Physical Graffiti.

The Eagles – certainly not a power band at the time.  Hotel California had fantastic velocity, the proponents lent them great veracity while volume and variety was building at the time.

Bob Seger – Night Moves had good velocity but not the same as Hotel California.  On the other hand Seger had great volume with 9 albums under his belt and variety switching from guitar to piano and other instruments.  His voice had a huge range and the proponents lent him much veracity because of this.

Rush – made it to #4 due to local popularity, being a Toronto Band and all.  They had some volume, releasing their 4th album, and veracity from local DJ’s, velocity was good, but again local. The variety was low so they only made to #4.

Scholarly Discussions on Semantic Search

David Amerland has a great grasp on Semantic Search and he did this great video below.  If you really want to know about Semantic Search then follow David on Google+ or buy his book – Google Semantic Search  – Here

 If you are interested in having me work on your website you can click through to my website page HarrisWeb Creative – Website Design or contact me directly at steve@harrisweb.ca

1 comment on “Understanding Semantic SearchAdd yours →

Leave a Reply

Your email address will not be published. Required fields are marked *