Not sure what size/kind of data you have, but it seems your problem is very similar to a recommender system task. Have you tried using matrix factorization method for recommender system?

Is this for a real-life application or a simplified model for exhibition purposes?

There's a ton of stuff to consider if this is for real. For example:

a) Donor available personal funds (donors having more to donate are more likely to donate more) [up to a saturation point]

b) Donor total donations (donors having donated more so far are more likely to donate more) [up to a saturation point]

c) Donation portfolia (donors donating to funds with certain properties/characteristics/features are more likely to donate to funds with similar properties/characteristics/features)

d) Donor social networks (donors who had social interaction with donors who donated to a fund are more likely to donate to said fund)

e) Donor fund satisfaction (if a donor has donated to a fund before, they are more likely to donate again if they are satisfied about how the fund handled their donation)

f) Fund recognition (funds with greater recogntion are more likely to receive donations)

g) Fund tax write off policies (funds with more convenient tax write off policies [for specific donors] are more likely to receive donations [from said donors])

h) ...

@Mohamed Baker Alawieh

I have tried recommender system using collaborative filtering. My dataset is something like this for example:

DONOR FUND SUM

D1 Fund X1 € 1000

D1 Fund X2 € 400

D2 Fund X10 € 5000

D1 Fund X3 € 100

D2 Fund X25 € 150

...

I used the SUM to provide the ratings for each donor. I calculated ratings something like RATING = SUM/max(SUM).

(I could also use a saturation point in calculating RATING as suggested by Asterios Chardalias )

So the dataset will look something like this for example:

DONOR FUND RATING

D1 Fund X1 1.00

D1 Fund X2 0.40

D2 Fund X10 1.00

D1 Fund X3 0.10

D2 Fund X25 0.03

...

The task is, since donor D1 has donated to Funds X1, X2, & X3; I want to calculate the likelihood of donor D1 donating to other Funds X4,X5,...,X100,..

If I use recommender system I get the predicted rating of donor D1 donating to other Funds but NOT the likelihood of D1 donating to other funds.

@ Asterios Chardalias

Yes it is a real life application.

I only have information about points b) & d).

I have the relationship info of donors but only for a few donors.

The only complete information I have is which donors donated for which fund & the fund amount.

The other info I have is about the PostCodes, Age, etc but has lots of missing values.

I do not have information about the properties of the funds, although I can do the research and hand-craft the properties of each fund.