Pattern Recognition

"You can't hide from the numbers" - Melissa Breen

Systematic Observation -

"the act of taking in raw data and making an action based on the 'category' of the pattern.." - Duda, Hart and Stork (2001)

Pattern recognition in itself, is a very broad topic. Real life examples as noted by Duda, Hart and Stork (2001) include:

- Identifying car keys in our pockets - Recognising a face

- Comprehending handwritten text

The very notion of Systematic Observation implies just that, it is systematic. What makes it so, is the need to have determined a conclusion before seeking the data. Working to a system does, in its very nature, give this activity a systematic approach.

When considering data's reliability, it must be noted that most data collected is statistically based; so having efficient gatherers of this data is a must. Data that is captured has a direct impact on whether pattern recognition is achievable and effective. With this in mind, it's validity is of paramount importance.

Supervised Learning -

"One set of observations (inputs) is assumed to be the cause of another set of observations (outputs)".

This is an area that I need to spend more time focussing on. In theory, it relates to all research however I would be very keen to dive further in to this over the course of the unit.

Learning Journey - needs work

Overall, the pattern recognition part of this unit is where I have struggled most. Getting my head around R will be the next step for the course as it will assist with things such as analysing large volumes of raw data.

I am familiar enough with recognising patterns in specific sports, but having the ability to conduct the research required to visualise these is the next step for me.

Examples of my understanding would be:

1. I would imagine there is a direct correlation between the amount of times a fielding side can bowl three consecutive maidens in a game of cricket, to when they next capture a wicket. I would suggest that the percentage of times they get a wicket in the next five overs would be extremely high, I look forward to building the tools to be able to research further in to this.

2. In AFL matches, I would assume that a teams Win-Loss Ratio is well above 0.5 if they ran at >80% efficiency with their disposal. Knowing where to find this information and how best to present it is what I am hoping to achieve across this semester.

#data