C++ code for partitioning data in partitions depending on their similarities.

Job Description:

I have two big files and I want to hash them to a hash table. The key is the line number and the value is a two dimensional array of the content in that line (from first file) and the content in that line (from second file).

1- min, max, avg, % data points that is a neighbor of some data points within a distance less than or equal to 2* average.

2- Partition the data points into groups/partition each is represented using a hash table where we try to group similar data points based on distances. Note that we need to allocate partition dynamically.

2.1 Pick 1st data point from k-NN table randomly.

2.2 Create a query Q to accept entries of form(point_ID, neighbor_point_ID,dist)

2.3 i=1

2.4 Create partition part_i

2.5 Pick randomly a point x from k-NN and insert x into the part_i

2.6 insert all x neighbors (x_ID,neighbor_point_ID,dist) in Q in order (based on their distances, the closest neighbor goes first).

2.7 x= “ ”

2.8 While the Q is not empty

2.8.1 Deque a data point x from the Q

2.8.2 If its distances from the original point to is less than 2 * avg then

Insert x it to part_i

Retrieve x k-NN from k-NN table ,and insert them to Q

Remove x from k-NN table

2.8.3 Else add x in separate list {x_ID,neighbor_point_ID,dist}, to be handled later

2.9 If Q is empty, then i=i+1 and create part_i go back to the while statement.

Aptitudini: Programare C++

ID Proiect: #28013786