How to apply BinCounts to weighted data

Question

For example, I have an array:

data = {1, 2, 3, 4, 5, 6, 7, 8, 9}

and a corresponding list of weighting factors

weight = {5, 4, 5, 2, 4, 2, 1, 1, 1}

Evaluating

BinCounts[data, {1, 9, 1}]

gives

{1, 1, 1, 1, 1, 1, 1, 1, 1},

but

{5, 4, 5, 2, 4, 2, 1, 1, 1}

is the result wanted. How can I get this?

BinCounts[ Catenate@MapThread[ConstantArray, {data, weight}], {1, 9, 1}] — ciao, Commented Jun 19, 2016 at 21:14
With WeightedData you must use Histogram or HistogramList. If the data points are all integers, use Merge[Total]@Thread[data -> weight]. — Szabolcs, Commented Jun 19, 2016 at 21:15
@ciao your solution works only with integer data and weights; could you please generalize it to non-integer data and weights? — Valerio, Commented Jun 3, 2017 at 17:24

Young · Accepted Answer · 2016-06-19 23:40:38Z

8

To formalize the comments, here is the complete code:

data = {1, 2, 3, 4, 5, 6, 7, 8, 9};
weight = {5, 4, 5, 2, 4, 2, 1, 1, 1};

Last[HistogramList[WeightedData[data, weight], 9]]

Which outputs

{5, 4, 5, 2, 4, 2, 1, 1, 1}

answered Jun 19, 2016 at 23:40

Young

7,5251 gold badge22 silver badges46 bronze badges

$\begingroup$ Thanks, Young. Can the weighting factor be negative? $\endgroup$
– dudulu
Commented Jun 20, 2016 at 2:28
$\begingroup$ No. I didn't know that was a requirement $\endgroup$
– Young
Commented Jun 20, 2016 at 2:30
$\begingroup$ The data can be negative. The weights and data can be fractions. $\endgroup$
– Young
Commented Jun 20, 2016 at 2:36
$\begingroup$ @dudulu, exactly what application is having you consider negative weights? $\endgroup$
– J. M.'s missing motivation ♦
Commented Jun 20, 2016 at 5:06
$\begingroup$ @Young your solution works well. However, it is quite slower than BinCounts. Could you please think of other solutions which may be faster? I need to apply this function to very large datasets. $\endgroup$
– Valerio
Commented Jun 3, 2017 at 17:26

Add a comment |

1 Answer 1