Today I learnt about decision trees. Decision trees are algorithms that classify things according to their features. To train the machine, a set of training data is used. Then, a set of unseen testing data is used to check to see if the algorithm works. What I did today was classify a group of monkeys into biting and non-biting monkeys by identifying each of their features. Then I wrote out a decision tree to test another set of monkeys to see if they were biting or not. To make the decision process faster, I put the most obvious features of a biting monkey at the top of the tree.
Monday 24 October 2022
Machine Learning
The task at hand was to, using a form of machine learning find an algorithm to separate biting monkeys from non-biting monkeys, to see which to put into a zoo and allow visitors to pet. The algorithm was to be based on attributes that were common between select monkeys in a certain category, including the position of the mouth, eyes, or accessories. In the end, I along with a fellow classmate was able to figure out a 100% working system to classify whether a monkey would bite or not, all except for one monkey which had a new type of mouth which meant there was not enough information to classify the monkey, landing it an uncertain spot in the biting category (you can't be too safe). Through this task, it was made semi-apparent to me how computers actually function within all the hardware, showcasing how important boolean information was to the processing of information, due to computers basically just being billions if not trillions of little bytes which only have two modes- on and off. This activity made it clear how computers functioned and sorted data, and how they do it with such accuracy and precision.
Categorising monkeys
From this classifying biting and non-biting monkeys task, I have developed a more precise idea and understanding of how computers classify information. The process of categorising starts with identifying patterns, in this case, the characteristics of the monkeys and the background colour. Then, classify the features into groups such as non-biting and biting monkeys. For me, I’ve found out that the majority of non-biting monkeys have green backgrounds and biting monkeys have blue backgrounds. Through this pattern, I am able to identify features that cause a monkey with a green background to be biting and vice versa and make a decision tree to categorise the monkeys. At last, I make an algorithm based on the decision tree and test it with other examples to further improve the algorithm. My algorithm worked and was able to accurately categorise the monkeys from the testing set.
Sunday 23 October 2022
Machine learning in Data Science
During our task today we learnt about the basics on Machine Learning:
- Machine learning is the application of Artificial Intelligence and give computers a set of data, allowing them to use the data to sort, filter and learn more through a given algorithm.
- As the Machine/computer is given more data sets, it becomes more accurate and more things can be added to the algorithm, improving the reliability and accuracy
Monkey - Biting or non-biting
The task that we did today requires us to create an algorithm to determine whether a monkey will bite or not. To do this task, we had to identify the right classifiers that can be used to create a decision tree. Some classifiers/factors that can be identified with each monkey include:
- Glasses
- Bowtie
- Mask
- Bandage
- Messy, non-messy hair
- Teeth
- Tongue stuck out
- One eye open/Both eyes closed/Both eyes open
- Smiling/sad
Using this classifiers we can make a decision tree and ask questions as all these are booleans (true or false)
For example, we can ask 'Is the monkey's mouth open?' If the answer is yes, we can go on to look at all the monkeys with their mouth open, look for more classifiers and ask more questions in order to filter it down
Once we have created a algorithm using the training set, we can use the testing data set to test whether the algorithm works and adjust accordingly
Machine Learning in Data Science 24/10/2022 - Aidan Ma 10E
Today in Data Science, I learnt about Machine Learning and Decision Trees. By identifying traits of the object you want to categorize, you can find patterns that they have. e.g. in this experiment we tried to see what traits biting monkeys and non-biting monkeys would have, in this experiment, the monkeys that have an open mouth are biting monkeys alongside large amounts of traits.
By making a list of what traits there are and checking for them, you can create a decision tree. By making sure that more obvious traits are on the top of the "tree" you can shorten the decision process.
Here's a list of traits for the monkeys:
Mouth Open
Glasses
Bowtie
Mask
Bandage
Messy Hair
Teeth
Tongue
Eyes Open
Eyes Closed
Smiling
One eye open
Machine Learning in data science
Today in data science, I learned how algorithms to sort and categorize things using specific characteristics, such as mouth eyes, bow, eye patches, and teeth.
The algorithm can use something called a decision tree to sort and classify the monkeys by biting and non-biting. AI can learn from the different characteristics if the monkey is biting or non-biting.
Mouth open
Glasses
Bow
Mask
Bandage
Messy Hair
Teeth
Tongue
Eyes
Eye brows
Smiling
One eye open
Both eyes open
Machine learning
Today I learnt that we need to use features to separate and analyse things in order to improve. and we also learnt many algorithms and we looked at monkeys. we took the features of monkeys like having mask close mouth , bow tie and dead have features that show if they bite or if they don't bite.
MONKEY
we use classifying and decision trees identify if a type of a monkey with a specific appearance will act in a certain way of biting or not biting.
we talked about the need for training in classification.
glasses
bowtie
mask
buandage
messy hair
teeth
tongue
eyes open
eyes brows
smiling one ye open
both eyes open
Subscribe to:
Posts (Atom)