Basic Algorithm Help

RichieD

Hi guys,

I wonder if anyone could any help me come up with a function for a program im writing?

So essentially I have 3 groups of customers,

1. Prefer to buy low priced Items
3. Prefer to buy high priced Items
2. Between 1 & 3

I then have a list of items which can have any price, so I need to work out a % chance of the customer buying the item.

Examples

Customer category 1 - Item price €1 = High chance to buy
Customer category 1 - Item price €100 = low chance to buy
Customer category 3 - Item price €1 = Low chance to buy
Customer category 3 - Item price €100 = High chance to buy

They dont need to be based on the actual amounts, just that the lower the price the more chance customer category 1 will buy it, the higher the price the higher the chance customer category 3 will buy it and 2 being in the middle ground.

Any help would be appreciated.

Thanks

Feathers

RichieD wrote: »

Hi guys,

I wonder if anyone could any help me come up with a function for a program im writing?

So essentially I have 3 groups of customers,

1. Prefer to buy low priced Items
3. Prefer to buy high priced Items
2. Between 1 & 3

I then have a list of items which can have any price, so I need to work out a % chance of the customer buying the item.

Examples

Customer category 1 - Item price €1 = High chance to buy
Customer category 1 - Item price €100 = low chance to buy
Customer category 3 - Item price €1 = Low chance to buy
Customer category 3 - Item price €100 = High chance to buy

They dont need to be based on the actual amounts, just that the lower the price the more chance customer category 1 will buy it, the higher the price the higher the chance customer category 3 will buy it and 2 being in the middle ground.

Any help would be appreciated.

Thanks

Not sure if it is as simple as this, but from what you described above, could you not do something like:

double likelyhoodToBuy;
switch(customerType) {
    case Customer.CHEAP:    likelyhoodToBuy = 100 - (itemPrice * (100/1)); 
                            break;
    case Customer.MODERATE: likelyhoodToBuy = (itemPrice * (100/1)) / 2;
                            break;
    case Customer.LAVISH:   likelyhoodToBuy =  itemPrice * (100/1);
                            break;
}

return likelyhoodToBuy;

Raging_Ninja

This seems like a bayesian analysis problem. The free online stanford AI course does something similar regarding spam detection, if you want to take a look you can find it here.