K Mean Clustering as a typical approach to Detectives
Before going forward, I would like you all to forget about K and Mean and just you focus on clustering for now and things will get clear on the way.
What is Clustering ?
Clustering is a key data science method, which does not require labeled data. That means clustering comes under Unsupervised Learning.
Everyone knows for the fact that cluster means a group of similar things or people positioned or occurring closely together.
The way we make judgements , clear our data , group the similar kind of data to the one side from rest is known as Clustering .
What is K Mean Cluster ?
K Mean term introduces a type of boundation to the way of grouping operation is performed . So here we basically perform Clustering in the following Steps:
Step 1: Choose the number of clusters k
This step can be either done manually or by making use of certain algorithms like Elbow Method.
This is a very important hyperparameter that needs to be decided very carefully other wise the model will be having certain errors and incomplete data.
Step 2: Select k random points from the data as centroids
This step involves initialization of mid point of k clusters. We can use some algorithm to decide the points or we may just re-assign them randomly since this K Mean Clustering is an Iterative approach so the real centroid will be found later.
For K = 2 , we have below scenario 👇
Step-3: Assign each data point to their closest centroid, which will form the predefined K clusters.
The distance of all the points are evaluated from the k centroid points and the nearest possible centroid is grouped as a part of Kth cluster.
For K = 2 , we have below scenario 👇
Step-4: Calculate the variance and place a new centroid of each cluster.
The Standard deviation or value of means of calculated for the kth cluster separately and the new centroid will be allocated in accordance with the mean.
Step-5: Repeat the third steps, which means reassign each data point to the new closest centroid of each cluster.
Once new centroid is evaluated , make the k clusters out of it in a similar fashion by assigning the points to their nearest centroid.
Step-6: Keep on Iterating over the cluster till you get the desired results.
The final cluster formed after this step would be best for satisfying all the needs of K clusters and thus we are able to separate the similar k features together.
How it is utilized by Detectives 🤔?
In the current generation, where’s there immense use of technology crime is taking over humanity , its not about only cyber crimes . In many countries or some major cities like Africa there have been cases of theft .
- Crimes are not limited to this. There are Banking frauds, Assaulting, Trafficking may even lead to Murder sometime 😶. Depending upon the type of Crime , we might be able to determine the probable location where a crime is bound to happen.
One thing why there would be any correlation 😕?
Yes , there would be correlation because you environment plays a vital role in determining the person you are. Although exceptions are everywhere but the nature of 85% in a locality remains the same.
- K Mean Clustering is also used in such a case to gather the suspects having similar properties like the accusation of crime and its relation with a people from certain location.
- It can be used as a way to train the newly joined forces in an area to have a pre loaded data of expectations of crime around that area. Thus preventing future incidents.
Here were some of the insights of K mean clustering and its relation in security field. Hope you enjoyed it 😊.