The process of the transformation of conceptual schema into a computerized form is called normalization. This is referred to as the database scheme.

The normalization process does not distinguish between an object set and a relationship. Therefore, it is referred to these by the relatively neutral term n-set. An n-set is a set of elements each of which is an n-vector. An n-set captures in itself the notion of an object set or a relationship as having n components. Each of these components will be called as a field.

There are different types of Normal Forms, They are…

**The first normal form (1NF):**

The first normal form relates to the structure of n-sets. It places a restriction on the nature of the fields involved in defining an n-set. This restriction is expressed is expresses in the following forms.

- The fields of n-set should have simple, atomic values
- n-sets should have no repeating groups.

In general, an n-set has to be split as many times as the number of non-atomic fields it contains. A representation of an n-set that is in 1NF is often called a normalized representation.

**The second normal form (2NF):**

The second normal form is concerned with eliminating redundancy in the structure of the representation of n-sets. There are three major problems that arise here.

- Update anomaly: This problem result in an inconsistency.
- Deletion anomaly: Deletion can result in loss of information. This is known as deletion anomaly and
- Insertion anomaly: An n-set is in the second normal form if it is in 1NF and each non-prime field is fully dependent upon each candidate key.

**Third normal form (3NF):**

The third normal form essentially tries to eliminate transitive dependencies from representations in 2NF. It is defined as follows:

An n-set is in third normal form if it is in 1NF and, for every collection C of fields, if any field not in C is functionally dependent on C, then all fields of the n-set are functionally dependent on C.

**Fourth normal form (4NF):**

Three of the above normal forms deal with functional dependencies only. It is possible for an n-set in 3NF to still exhibit update, insertion and deletion anomalies.

In order to define 4NF, one has to define trivial multivalued dependencies. A trivial multivalued dependency is one which necessarily holds for any vector. For example, X->0, where 0 is the null list, always holds for any vector. Likewise, X->-> Y holds for any vector consisting of X and Y only. That is, if there is an n-set consisting of fields of fields A1, B1 and C1, then

{A1, B1} -> -> C1

An n-set is in fourth normal form, if whenever a non trivial multivalued dependency X->-> Y holds for an n-set then so does the functional dependency Y-> A hold for every field A belonging to the n-set.