0
$\begingroup$

I have a regression problem. When a truck comes, it influences the demand of employees for the next 30 days. Additionally the demand depends on the type of truck (when the truck is big, we need more people).

Which algorithm/model can help to predict a demand on employees on the defined day?

Data looks like so (I cannot give the original data):

|-----------|--------|----------------|-------|------------------|----------|
|Transaction|Employee|Date_Transaction| Truck |Arrival_Date_Truck|Type_Truck|
|-----------|--------|----------------|-------|------------------|----------|
|     1     |   A    | 01.01.2010     |Truck_B|     07.12.2009   |    Big   |
|-----------|--------|----------------|-------|------------------|----------|
|     2     |   B    | 01.01.2010     |Truck_A|     05.12.2009   |    Big   |
|-----------|--------|----------------|-------|------------------|----------|
|     3     |   A    | 02.01.2010     |Truck_A|     05.12.2009   |   Small  |
|-----------|--------|----------------|-------|------------------|----------|
|     4     |   C    | 02.01.2010     |Truck_B|     07.12.2009   |   Small  |
|-----------|--------|----------------|-------|------------------|----------|
|     5     |   A    | 03.01.2010     |Truck_C|     12.12.2009   |   Middle |
|-----------|--------|----------------|-------|------------------|----------|
|     6     |   B    | 03.01.2010     |Truck_C|     12.12.2009   |   Middle |
|-----------|--------|----------------|-------|------------------|----------|
|     7     |   C    | 03.01.2010     |Truck_B|     07.12.2009   |    Big   |
|-----------|--------|----------------|-------|------------------|----------|
|     8     |   D    | 03.01.2010     |Truck_B|     07.12.2009   |    Big   |
|-----------|--------|----------------|-------|------------------|----------|
|     9     |   B    | 04.01.2010     |Truck_C|     12.12.2009   |   Middle |
|-----------|--------|----------------|-------|------------------|----------|

I know, that a count of days (count of transactions), that truck needs depends on the type of truck. Furthermore I know, that the count of transactions looks like: distribution

$\endgroup$
0
$\begingroup$

Any regression algorithm can address this problem, to a greater or lesser degree of success- the important thing is in how you form your data.

One reasonable form to use would be a structure with one row for each day and features describing the arrival of trucks over an appropriate timeframe- say, over the past 14 days. Each row would then also have the demand for the day associated with it as the regression target.

You would then be able to train a model that predicts the demand for a given day provided the appropriate truck arrival history.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.