One nearest neighbor using awk

[Murlidhar]

This is what I am trying to do using AWK language. I have problem with mainly step 2. I have shown a sample dataset but the original dataset consists of 100 fields and 2000 records.

Algorithm

1) initialize accuracy = 0
[pre]2) for each record r
--Find the closest other record, o, in the dataset using distance formula[/pre]
3) if the class value of closest record o is equal to class value of current record, increment accuracy by 1. here, class value is last field(col 6)
4) Finally, 100 * accuracy/total_records

Sample Dataset

[pre] c1 c2 c3 c4 c5 c6 --> Columns
0.6 0.1 0.2 0.3 0.4 0.3 --> row1 & row7 nearest neighbor in c1
0.1 0.2 0.1 0.1 0.1 0.6 and same values in c6(0.3) so ++accuracy
0.2 0.3 0.1 0.1 0.2 0.6
0.3 0.4 0.1 0.1 0.3 0.3
0.4 0.5 0.1 0.1 0.9 0.6
0.5 0.6 0.1 0.1 0.8 0.9
0.6 0.7 0.1 0.1 0.7 0.3
0.7 0.8 0.1 0.1 0.6 0.6
0.8 0.9 0.1 0.1 0.5 0.9
0.9 1.0 0.1 0.1 0.4 0.3

[/pre]Code

[pre]BEGIN{
accuracy = 0;
total_records = 10;
}
{
for(i = 1; i <= 5; i++ ) # for fields 1 to 5 only
#for each record
{
#find closest record(calculating the distance)
distance = abs($i - other_records)

#compare values of field 6 for closest and current(each) record
if(current_record_field_6.value == closest_record_field_6.value)
{
++accuracy;
}
}
}
END{
percentage = 100 * (accuracy/total_records);

print percentage;
}
[/pre]I am struggling on how to find the closest record for each record in the dataset using AWK. As far as I know '{}' block is only executed once for each record.

Any help or suggestion is much appreciated.

 

[mikrom]

I am struggling on how to find the closest record for each record in the dataset using awk. As far as I know '{}' block is only executed once for each record.

Yes block {..} is executed only once for each record. In this block we can store every record in an array, then compute the distances of every record with all previous records and store the values in a distance matrix.

Then in the block END{..} we can evaluate computed distances and print the results.

I tried it. For the input data posted above I computed this distance matrix (rN means N-th record)

REC  r01   r02   r03   r04   r05   r06   r03   r08   r09   r10 
r01 0.00  1.20  1.10  1.00  1.40  1.30  1.20  1.30  1.40  1.50 
r02 1.20  0.00  0.30  0.60  1.40  1.50  1.60  1.70  1.80  1.90 
r03 1.10  0.30  0.00  0.30  1.10  1.20  1.30  1.40  1.50  1.60 
r04 1.00  0.60  0.30  0.00  0.80  0.90  1.00  1.10  1.20  1.30 
r05 1.40  1.40  1.10  0.80  0.00  0.30  0.60  0.90  1.20  1.50 
r06 1.30  1.50  1.20  0.90  0.30  0.00  0.30  0.60  0.90  1.20 
r07 1.20  1.60  1.30  1.00  0.60  0.30  0.00  0.30  0.60  0.90 
r08 1.30  1.70  1.40  1.10  0.90  0.60  0.30  0.00  0.30  0.60 
r09 1.40  1.80  1.50  1.20  1.20  0.90  0.60  0.30  0.00  0.30 
r10 1.50  1.90  1.60  1.30  1.50  1.20  0.90  0.60  0.30  0.00

and I got these results:

record #01: [ 0.6 0.1 0.2 0.3 0.4 0.3 ]
	closest record(s) with minimal distance =  1.00
		record #04: [ 0.3 0.4 0.1 0.1 0.3 0.3 ]	--> accuracy found

record #02: [ 0.1 0.2 0.1 0.1 0.1 0.6 ]
	closest record(s) with minimal distance =  0.30
		record #03: [ 0.2 0.3 0.1 0.1 0.2 0.6 ]	--> accuracy found

record #03: [ 0.2 0.3 0.1 0.1 0.2 0.6 ]
	closest record(s) with minimal distance =  0.30
		record #02: [ 0.1 0.2 0.1 0.1 0.1 0.6 ]	--> accuracy found
		record #04: [ 0.3 0.4 0.1 0.1 0.3 0.3 ]

record #04: [ 0.3 0.4 0.1 0.1 0.3 0.3 ]
	closest record(s) with minimal distance =  0.30
		record #03: [ 0.2 0.3 0.1 0.1 0.2 0.6 ]

record #05: [ 0.4 0.5 0.1 0.1 0.9 0.6 ]
	closest record(s) with minimal distance =  0.30
		record #06: [ 0.5 0.6 0.1 0.1 0.8 0.9 ]

record #06: [ 0.5 0.6 0.1 0.1 0.8 0.9 ]
	closest record(s) with minimal distance =  0.30
		record #05: [ 0.4 0.5 0.1 0.1 0.9 0.6 ]
		record #07: [ 0.6 0.7 0.1 0.1 0.7 0.3 ]

record #07: [ 0.6 0.7 0.1 0.1 0.7 0.3 ]
	closest record(s) with minimal distance =  0.30
		record #06: [ 0.5 0.6 0.1 0.1 0.8 0.9 ]
		record #08: [ 0.7 0.8 0.1 0.1 0.6 0.6 ]

record #08: [ 0.7 0.8 0.1 0.1 0.6 0.6 ]
	closest record(s) with minimal distance =  0.30
		record #07: [ 0.6 0.7 0.1 0.1 0.7 0.3 ]
		record #09: [ 0.8 0.9 0.1 0.1 0.5 0.9 ]

record #09: [ 0.8 0.9 0.1 0.1 0.5 0.9 ]
	closest record(s) with minimal distance =  0.30
		record #08: [ 0.7 0.8 0.1 0.1 0.6 0.6 ]
		record #10: [ 0.9 1.0 0.1 0.1 0.4 0.3 ]

record #10: [ 0.9 1.0 0.1 0.1 0.4 0.3 ]
	closest record(s) with minimal distance =  0.30
		record #09: [ 0.8 0.9 0.1 0.1 0.5 0.9 ]

number of records processed: 10
percentage = 30