You are on page 1of 3

MINNIE 

Machine Learning Report 


 
Describe what Binary Tree is and how machine learning works 
(Data, Convert, Train, Predict). Add a diagram picture.  
 
A  binary  tree  is  a  tree  data  structure  with  one  root  that  branches  up 
to  two  children  nodes,  each  children  nodes  replicating  itself  over  and  over. 
The  topmost node is called the root. The branched nodes are called children 
nodes  to  the  parent  nodes  or  nodes  where  they  branch  from.  The  root 
contains  conditions  which  filters  whether  a  test  subject  is  true  or  false 
according  to  those  conditions.  The  next  children  nodes  contains  more 
specific  conditions which allow the AI to narrow down the answer to just one 
class. 
The  basic  process  starts  from  providing  data  for  the  AI.  This  is  done 
on  Google  Sheets  which  contains  the  information  of  each  class  and  its 
qualities.  Then,  that  data  needs  to  be  converted  into  numeral  format,  so 
that  the  AI  can  understand  it.  After  that,  some  more  data  needs  to  be 
inputted  in  order  to  train  the  AI.  Then,  the  AI  would  be  ready  to  predict 
answers based on the data we’ve given it. 

 
MINNIE 

Describe your topic & data with a picture or chart. What mistakes 
or challenges did you face with this data?  
 
I  chose  for  the  AI  to  be  about  classifying  boiled  eggs  into  three 
categories:  poached,  soft-boiled  and hard-boiled. As they are closely related 
and  visually  similar,  they  share  quantifiable  characteristics  like  the  white’s 
opacity, the yolk’s state, the time boiled and the shape. 
At first I was confused to what I was supposed to make. I was 
struggling to understand that the concept of binary tree is similar to 
akinator and that there needs to be exactly two definite answers for each 
quality’s condition. But once I understood that part, making the data and 
converting the data into numbers is not so difficult.  

 
 
MINNIE 

How accurate is your AI? Why? Explain with pictures of your results 
 
My AI is quite accurate as the data is definite. There’s also not a lot 
of options for each quality(the white, yolk, boil time and shape). It is quite 
probable that if the test subject fits this particular condition, it will 
automatically be classified. For example, if the test subject’s white is 
translucent, it is automatically a poached egg and not the other two. The 
only quality here that varies the most is the boil time. If the boil time of the 
test data becomes more specific, as in become decimals, the AI might have 
some problems to process numbers in between 3 and 4, 5 and 6. 

You might also like