WHAT IS ASSOCIATION ANALYSIS AND TERMINOLOGIES OF ASSOCIATION ANALYSIS

WHAT IS ASSOCIATION ANALYSIS?

Data science technology is the most influential technology of the present time. The first step of this technology is analyzing the data. One of them is the association analysis. Association analysis is used to recognize hidden relationships between the data. The relationships between the data are very useful. Data analysts can extract any information from it. The relationships which are recognized from the data are expressed as the compilation of association rules. Association rules do not find the relationships between individuals, it recognizes relationships among the group of the people. For understanding association analysis and association rules better, we will cite an example of a supermarket where all the relevant items are grouped together.

COMPONENTS OF ASSOCIATION RULES

There are two components of association rules. They are listed below-->
• Antecedent
• Consequent

Antecedent and consequent are separate lists of the items (kept in a supermarket). There is a term called item set. Item set is the union of all the items of consequent and antecedent. For example, in the antecedent item list, there are tomato and potato. On the other hand, in the consequent item list, there are egg and bread. The item set will be potato, tomato, egg, and bread.

TERMINOLOGIES OF ASSOCIATION ANALYSIS

Here are some factors, terms, etc. which are frequently used in the association analysis. They are listed below-->
• Support
• Confidence
• Lift


WHAT IS SUPPORT?

Support is a terminology which shows the maximum occurrence of an item. This would be clearer if we cite an example. Let us say we have an item list1 which contain the only item that is a toothbrush. On the other hand, we have item list2 in which we have one item that is butter. Now, we will check the transactions of the customers. Here transactions mean which item has been sold the most. In our case, let us say people have purchased more butter as compared to the toothbrush. The butter is present in item list2. Then, we say that item list2 has higher support. There can be one more case where people who have purchased both the items. We can express support mathematically as the ratio of transactions of both the lists and the total number of transactions.

WHAT IS CONFIDENCE?

Confidence is a terminology which shows the probability of purchasing an item (consequent) after purchasing an item (antecedent). This would be clearer when we cite an example. Let us say, there is an antecedent list, which contains item bread in it. A customer has purchased bread. Then we can predict that the customer will also purchase butter. There is a probability that a person can purchase butter too. Then we say that purchasing butter has high confidence. On the other hand, if a customer has purchased bread, then there is a very low probability that he will purchase a toothbrush. In this case, we say that the purchasing toothbrush has low confidence.

CONCLUSION

Association analysis is a very fascinating topic of data science technology. Those who are interested in learning association analysis can enroll themselves here for a Data Science Course in Pune.