Page 279 - AI Ver 3.0 Class 11
P. 279

AI data analysis employs AI techniques and data science to enhance the processes of cleaning, inspecting, and
                 modelling over both structured and unstructured data. The main goal is to extract valuable information that can aid
                 in decision-making and drawing conclusions.


                         Data Collection


                 Data collection lets you record past events, so you can use data analysis
                 to find patterns. From these patterns, you can build predictive models
                 using machine learning to spot trends and predict future changes.
                 Data collection means gathering data from many sources, both offline
                 and online. Collecting large amount of data can be the hardest part of a
                 machine learning project, especially on a large scale. The amount of data
                 you need depends on the number of features in the data set. It’s best to
                 collect as much data as possible for accurate predictions. You can start
                 with small batches of data and see how the model performs. It’s important             DATA
                 to collect diverse data to ensure your model covers various scenarios.
                 The quantity of data also depends on, how complex your model is. For
                 simple tasks like, license plate detection, small batches of data might be
                 enough. But for more complex tasks like, medical AI, you will need a lot of data. Before collecting data, data scientists
                 need to understand the problem, the best solution, and the data requirements. Based on these requirements, they
                 identify the data sources and collect the data. Data is essential for any project, and it is needed throughout the
                 project’s development. Therefore, identifying data needs, collecting data, and analysing it repeatedly.

                 Data is the main requirement for data collection methods. In any type of research or company operations, collecting
                 data serves primarily to support the identification of important variables, performance being the most important
                 among them. Thus, the process of collecting data plays a vital role almost in every field. Depending on the kind of
                 data being collected, there are two main categories of data collection methods: primary and secondary.

                 Primary Data Source
                 A primary data source refers to the original source from which data is collected firsthand. This data is obtained
                 directly from its origin, without any intermediary sources or interpretations. Primary data sources include surveys,
                 interviews, observations, experiments, and any other method where data is collected directly by the researcher or
                 organisation for a specific purpose. This type of data is considered valuable because it is tailored as per the specific
                 research or business needs and is often more accurate and relevant than secondary data, which is obtained from
                 sources which have already interpreted or analysed the original data.


                                                             Primary Data Sources




                                                                                          Marketing
                      Survey         Interview       Observation        Experiment                         Questionnaire
                                                                                           Campaign
                 Let us study each source.







                                                                   Data Literacy—Data Collection to Data Analysis  277
   274   275   276   277   278   279   280   281   282   283   284