Reputation: 314
I am a starter in the field of ML and wants to understand "Multi Dimensional Array". My current understanding is as follows. Please help me where i am wrong.
Suppose I have data of customers, we can refer customers as items.
Now suppose each item i.e. each individual has a Customer identification no (CIN) - we can call this as a one dimensional matrix.
extending the same example,
now each item has list of attributes e.g. CIN, Gender, Height, weight, Eye colour. If i want to arrange them it will be a 2 dimensional matrix. Row representing students, columns representing their physical attributes "Roll No, Gender, Height, weight, Eye colour".
Now suppose i add another list of different attributes - Past purchase history i.e. Last purchase date and invoice no.
Can we call a matrix which has both phyical attributes "Roll No, Gender, Height, weight, Eye colour" and "Last purchase date and invoice no." is this a 3 dimensional matrix ?
Is it the grouping into physical and purchase history which will make it a 3 d matrix.
I can still store it as a 2 d matrix (like a spread sheet) with physical and purchase history as columns or like 2 sheets with common link as CIN.
Upvotes: 1
Views: 66
Reputation: 1481
I think you are getting confused between dimensions of an array and dimensions for a ML model also called features or independent variables.
When you train a model each piece of information that you pass to the model is a feature or a dimension. So for example Roll No, Gender, Height are all features. If you add Last purchase date and invoice those are also features. If you add all of them in a dataframe, each row of the dataframe can be thought of as an 1-D array.
The grouping of physical attributes and purchase history which will not make it a 3-D matrix.
3-D arrays or n-D arrays are called tensors. They are something completely different.
Upvotes: 1