I have written a program that defines a relation between a 'full' data table a set of sets of column headers (features), an aggregate function and the aggregate table.
An example query :
?- data(D), fulltable_aggfunction_sets_aggtable(D,mean,[[a,b,c],[d,e,f]],AggTable),
print_data(D),print_data(AggTable).
[a,b,c,d,e,f,g,class]
[1,1,1,1,1,1,1,0]
[2,3,4,5,4,2,1,0]
[3,1,3,4,6,7,8,1]
[1,2,3,6,1,1,2,1]
[feature(1,mean,[a,b,c]),feature(2,mean,[d,e,f])]
[1,1]
[3,3.6666666666666665]
[2.3333333333333335,5.666666666666667]
[2,2.6666666666666665]
D = [[a, b, c, d, e, f, g, class], [1, 1, 1, 1, 1, 1, 1|...], [2, 3, 4, 5, 4, 2|...], [3, 1, 3, 4, 6|...], [1, 2, 3, 6|...]],
AggTable = [[feature(1, mean, [a, b, c]), feature(2, mean, [d, e, f])], [1, 1], [3, 3.6666666666666665], [2.3333333333333335, 5.666666666666667], [2, 2.6666666666666665]]
The following is my code:
fulltable_aggfunction_sets_aggtable(Full,Func,Sets,Aggtable):-
Full =[Features|Data],
flist_sets_indexs(Features,Sets,Indexs),
maplist(indexs_flist_sets(Indexs),Data,Datasplits),
maplist(aggfun_listoflists_values(Func),Datasplits,AggData),
list_indexes(Sets,SetIndex),
maplist(name_set_id_feature(Func),Sets,SetIndex,FeatureNames),
append([FeatureNames],AggData,Aggtable).
name_set_id_feature(Func,Set,Id,feature(Id,Func,Set)).
list_indexes(List,Indexes):-
findall(I,nth1(I,List,_),Indexes).
aggfunc_list_value(sum,List,Value):-
sumlist(List,Value).
aggfunc_list_value(mean,List,Value):-
sumlist(List,Sum),
length(List,L),
Value is Sum/L.
aggfun_listoflists_values(Fun,ListsofLists,Values):-
maplist(aggfunc_list_value(Fun),ListsofLists,Values).
my_nth0(List,Elem,I):- nth0(I,List,Elem).
indexs_flist_sets(I,F,S):-flist_sets_indexs(F,S,I).
flist_sets_indexs(Features,Sets,Indexs):-
maplist(flist_set_indexes(Features),Sets,Indexs).
flist_set_indexes(Features,Set,Indexs):-
maplist(my_nth0(Features),Set,Indexs).
%aux
print_data(Data_set):-
maplist(print_line,Data_set).
print_line(Data_line):-
format("~w ~n",[Data_line]).
data(Data):-
Data =[[a,b,c,d,e,f,g,class],
[1,1,1,1,1,1,1,0],
[2,3,4,5,4,2,1,0],
[3,1,3,4,6,7,8,1],
[1,2,3,6,1,1,2,1]].
In my code I have these two lines which just reorder arguments so that I can pass these rules into a maplist.
my_nth0(List,Elem,I):- nth0(I,List,Elem).
indexs_flist_sets(I,F,S):-flist_sets_indexs(F,S,I).
Is there a better way to do this? How can I use maplist in this case so as not to have to define these rules?
in SWI-Prolog there are library(yall) and library(lambda).
Using library(yall), for instance
flist_set_indexes(Features,Set,Indexs):-
maplist({Features}/[Elem,I]>>nth0(I,Features,Elem),Set,Indexs).
print_data(Data_set):-
maplist([Data_line]>>format("~w ~n",[Data_line]),Data_set).
while using the second
:- use_module(library(lambda)).
flist_set_indexes(Features,Set,Indexs):-
maplist(\Elem^I^nth0(I,Features,Elem),Set,Indexs).
%aux
print_data(Data_set):-
maplist(\Data_line^format("~w ~n",[Data_line]),Data_set).
Libraries have similar functionalities. but note that yall requires a 'capture' of Features. library(lambda) is available after
?- pack_install(lambda).
while library(yall) is autoloaded