Page 307 - Ai_V3.0_c11_flipbook
P. 307

For example:
                                                                  15 3 
                                                             A =         k3
                                                                           =
                                                                  2 47 
                                                             1 3 53 33 ×     3 15 9 
                                                                  ×
                                                             ×
                                                     =
                                                   kA 3A =                ⇒        
                                                              ×
                                                                       ×
                                                                  ×
                                                             2 3 4 3 7 3     6 12 21 
                        Applications of Matrices in AI
                 Matrices are used throughout the field of machine learning for computing:
                    • Sales Forecasting and Price Prediction: Matrices are used to represent relevant predictor and response variables.
                    • Image Processing: Digital images can be represented using matrices. In the following figure, each box appears
                   transparent or coloured. It is represented in the corresponding matrix with either a “1” or a “0”. A zero means the box
                   contains nothing and is white, while a one indicates that the box is filled (in this case it is filled with black).


                                                                           0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
                                                                           0  1  1  1  0  0  0  0  0  0  0  0  0  1  1  1  0
                                                                           0  1  1  1  1  0  0  0  0  0  0  0  1  1  1  1  0
                                                                           0  1  1  0  1  1  0  0  0  0  0  1  1  0  1  1  0
                                                                           0  1  1  0  0  1  1  0  0  0  1  1  0  0  1  1  0
                                                                           0  1  1  0  0  0  1  1  1  1  1  0  0  0  1  1  0
                                                                           0  1  1  0  0  0  0  1  1  1  0  0  0  0  1  1  0
                                                                           0  1  1  0  0  0  0  0  0  0  0  0  0  0  1  1  0
                                                                           0  1  1  0  0  0  0  0  0  0  0  0  0  0  1  1  0
                                                                           0  1  1  0  0  0  0  0  0  0  0  0  0  0  1  1  0
                                                                           0  1  1  0  0  0  0  0  0  0  0  0  0  0  1  1  0
                                                                           0  1  1  0  0  0  0  0  0  0  0  0  0  0  1  1  0
                                                                           0  1  1  0  0  0  0  0  0  0  0  0  0  0  1  1  0
                                                                           0  1  1  0  0  0  0  0  0  0  0  0  0  0  1  1  0
                                                                           0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0


                    • Recommender Systems: They use matrices to relate between users' purchased and users' viewed item(s).
                    • Natural Language Processing: In NLP, vectors represent the distribution of a particular word in a document. Vectors
                   are one-dimensional matrices.


                        Data Preprocessing


                 Data preprocessing is an essential phase in the machine learning process that prepares datasets for effective machine
                 learning applications. It includes multiple processes to clean, transform, reduce, integrate, and normalise data:

                           Data               Data             Data Reduction      Data Integration       Feature
                          Cleaning        Transformation                           & Normalisation       Selection


                 1.  Data Cleaning: Businesses have abundance of data. However, not all of it is accurate
                   or organised. When it comes to machine learning, if data is not sufficiently cleaned, the
                   accuracy of your model is at risk. The following steps are taken to clean/prepare the
                   data:
                    •  Missing Data: Missing data refers to the absence of certain values in the dataset,
                      which can result from various causes. To handle missing data, strategies include
                      removing rows or columns with missing values, imputing missing values with
                      estimates, or utilising algorithms that can manage missing data.

                                                                   Data Literacy—Data Collection to Data Analysis  305
   302   303   304   305   306   307   308   309   310   311   312