I have N objects, and M sets of those objects. Sets are non-empty, different, and may intersect. Typically M and N are of the same order of magnitude, usually M > N.
Historically my sets were encoded as-is, each just contained a table (array) of its objects, but I'd like to create a more optimized encoding. Typically some objects present in most of the sets, and I want to utilize this.
My idea is to represent sets as stacks (i.e. single-directional linked lists), whereas their bottom parts can be shared across different sets. It can also be defined as a tree, whereas each node/leaf has a pointer to its parent, but not children.
Such a data structure will allow to use the most common subsets of objects as roots, which all the appropriate sets may "inherit".
The most efficient encoding is computed by the following algorithm. I'll write it as a recursive pseudo-code.
BuildAllChains()
{
BuildSubChains(allSets, NULL);
}
BuildSubChains(sets, pParent)
{
if (sets is empty)
return;
trgObj = the most frequent object from sets;
pNode = new Node;
pNode->Object = trgObj;
pNode->pParent = pParent;
newSets = empty;
for (each set in sets that contains the trgObj)
{
remove trgObj from set;
remove set from sets;
if (set is empty)
set->pHead = pNode;
else
newSets.Insert(set);
}
BuildSubChains(sets, pParent);
BuildSubChains(newSets, pNode);
}
Note: the pseudo-code is written in a recursive manner, but technically naive recursion should not be used, because at each point the splitting is not balanced, and in a degenerate case (which is likely, since the source data isn't random) the recursion depth would be O(N).
Practically I use a combination of loop + recursion, whereas recursion always invoked on a smaller part.
So, the idea is to select each time the most common object, create a "subset" which inherits its parent subset, and all the sets that include it, as well as all the predecessors selected so far - should be based on this subset.
Now, I'm trying to figure-out an effective way to select the most frequent object from the sets. Initially my idea was to compute the histogram of all the objects, and sort it once. Then, during the recursion, whenever we remove an object and select only sets that contain/don't contain it - deduce the sorted histogram of the remaining sets. But then I realized that this is not trivial, because we remove many sets, each containing many objects.
Of course we can select each time the most frequent object directly, i.e. O(N*M). But it also looks inferior, in a degenerate case, where an object exists in either almost all or almost none sets we may need to repeat this O(N) times. OTOH for those specific cases in-place adjustment of the sorted histogram may be preferred way to go.
So far I couldn't come up with a good enough solution. Any ideas would be appreciated. Thanks in advance.
Update:
#Ivan: first thanks a lot for the answer and the detailed analysis.
I do store the list of elements within the histogram rather than the count only. Actually I use pretty sophisticated data structures (not related to STL) with intrusive containers, corss-linked pointers and etc. I planned this from the beginning, because than it seemed to me that the histogram adjustment after removing elements would be trivial.
I think the main point of your suggestion, which I didn't figure-out myself, is that at each step the histograms should only contain elements that are still present in the family, i.e. they must not contain zeroes. I thought that in cases where the splitting is very uneven creating a new histogram for the smaller part is too expensive. But restricting it to only existing elements is a really good idea.
So we remove sets of the smaller family, adjust the "big" histogram and build the "small" one. Now, I need some clarifications about how to keep the big histogram sorted.
One idea, which I thought about first, was immediate fix of the histogram after every single element removal. I.e. for every set we remove, for every object in the set, remove it from the histogram, and if the sort is broken - swap the histogram element with its neighbor until the sort is restored.
This seems good if we remove small number of objects, we don't need to traverse the whole histogram, we do a "micro-bubble" sort.
However when removing large number of objects it seems better to just remove all the objects and then re-sort the array via quick-sort.
So, do you have a better idea regarding this?
Update2:
I think about the following: The histogram should be a data structure which is a binary search tree (auto-balanced of course), whereas each element of the tree contains the appropriate object ID and the list of the sets it belongs to (so far). The comparison criteria is the size of this list.
Each set should contain the list of objects it contains now, whereas the "object" has the direct pointer to the element histogram. In addition each set should contain the number of objects matched so far, set to 0 at the beginning.
Technically we need a cross-linked list node, i.e. a structure that exists in 2 linked lists simultaneously: in the list of a histogram element, and in the list of the set. This node also should contain pointers to both the histogram item and the set. I call it a "cross-link".
Picking the most frequent object is just finding the maximum in the tree.
Adjusting such a histogram is O(M log(N)), whereas M is the number of elements that are currently affected, which is smaller than N if only a little number is affected.
And I'll also use your idea to build the smaller histogram and adjust the bigger.
Sounds right?
I denote the total size of sets with T. The solution I present works in time O(T log T log N).
For the clarity I denote with set the initial sets and with family the set of these sets.
Indeed, let's store a histogram. In BuildSubChains function we maintain a histogram of all elements which are presented in the sets at the moment, sorted by frequency. It may be something like std::set of pairs (frequency, value), maybe with cross-references so you could find an element by value. Now taking the most frequent element is straightforward: it is the first element in the histogram. However, maintaining it is trickier.
You split your family of sets into two subfamilies, one containing the most frequent element, one not. Let there total sizes be T' and T''. Take the family with the smallest total size and remove all elements from its sets from the histogram, making the new histogram on the run. Now you have a histogram for both families, and it is built in time O(min(T', T'') log n), where log n comes from operations with std::set.
At the first glance it seems that it works in quadratic time. However, it is faster. Take a look at any single element. Every time we explicitly remove this element from the histogram the size of its family at least halves, so each element will directly participate in no more than log T removals. So there will be O(T log T) operations with histograms in total.
There might be a better solution if I knew the total size of sets. However, no solution can be faster than O(T), and this is only logarithmically slower.
There may be one more improvement: if you store in the histogram not only elements and frequencies, but also the sets that contain the element (simply another std::set for each element) you'll be able to efficiently select all sets that contain the most frequent element.
Related
Apparently you could do either, but the former is more common.
Why would you choose the latter and how does it work?
I read this: http://www.drdobbs.com/cpp/stls-red-black-trees/184410531; which made me think that they did it. It says:
insert_always is a status variable that tells rb_tree whether multiple instances of the same key value are allowed. This variable is set by the constructor and is used by the STL to distinguish between set and multiset and between map and multimap. set and map can only have one occurrence of a particular key, whereas multiset and multimap can have multiple occurrences.
Although now i think it doesnt necessarily mean that. They might still be using containers.
I'm thinking all the nodes with the same key would have to be in a row, because you either have to store all nodes with the same key on the right side or the left side. So if you store equal nodes to the right and insert 1000 1s and one 2, you'd basically have a linked list, which would ruin the properties of the red black tree.
Is the reason why i can't find much on it that it's just a bad idea?
down side of store as multiple nodes:
expands tree size, which make search slower.
if you want to retrieve all values for key K, you need M*log(N) time, where N is number of total nodes, M is number of values for key K, unless you introduce extra code (which complicates the data structure) to implement linked list for these values. (if storing collection, time complexity only take log(N), and it's simple to implement)
more costly to delete. with multi-node method, you'll need to remove node on every delete, but with collection-storage, you only need to remove node K when the last value of key K is deleted.
Can't think of any good side of multi-node method.
Binary Search trees by definition cannot contain duplicates. If you use them to produce a sorted list throwing out the duplicates would produce an incorrect result.
I am working on an implementation of Red Black trees in PHP when I ran into the duplicate issue. We are going to use the tree for sorting and searching.
I am considering adding an occurrence value to the node data type. When a duplicate is encountered just increment occurrence. When walking the tree to produce output just repeat the value by the number of occurrences. I think I would still have a valid BST and avoid having a whole chain of duplicate values which preserve the optimal search time.
I need to calculate a 1d histogram that must be dynamically maintained and looked up frequently. One idea I had involves keeping an ordered array with the data (cause thus I can determine percentiles in O(1), and this suffices for quickly finding a histogram with non-uniform bins with the exactly same amount of points inside each bin).
So, is there a way that is less than O(N) to insert a number into an ordered array while keeping it ordered?
I guess the answer is very well known but I don't know a lot about algorithms (physicists doing numerical calculations rarely do).
In the general case, you could use a more flexible tree-like data structure. This would allow access, insertion and deletion in O(log) time and is also relatively easy to get ready-made from a library (ex.: C++'s STL map).
(Or a hash map...)
An ordered array with binary search does the same things as a tree, but is more rigid. It might probably be faster for acess and memory use but you will pay when having to insert or delete things in the middle (O(n) cost).
Note, however, that an ordered array might be enough for you: if your data points are often the same, you can mantain a list of pairs {key, count}, ordered by key, being able to quickly add another instance of an existing item (but still having to do more work to add a new item)
You could use binary search. This is O(log(n)).
If you like to insert number x, then take the number in the middle of your array and compare it to x. if x is smaller then then take the number in the middle of the first half else the number in the middle of the second half and so on.
You can perform insertions in O(1) time if you rearrange your array as a bunch of linked-lists hanging off of each element:
keys = Array([0][1][2][3][4]......)
a c b e f . .
d g i . . .
h j .
|__|__|__|__|__|__|__/linked lists
There's also the strategy of keeping two datastructures at the same time, if your update workload supports it without increasing time-complexity of common operations.
So, is there a way that is less than O(N) to insert a number into an
ordered array while keeping it ordered?
Yes, you can use an array to implement a binary search tree using arrays and do the insertion in O(log n) time. How?
Keep index 0 empty; index 1 = root; if node is the left child of parent node, index of node = 2 * index of parent node; if node is the right child of parent node, index of node = 2 * index of parent node + 1.
Insertion will thus be O(log n). Unfortunately, you might notice that the binary search tree for an ordered list might degenerate to a linear search if you don't balance the tree i.e. O(n), which is pointless. Here, you may have to implement a red black tree to keep the height balanced. However, this is quite complicated, BUT insertion can be done with arrays in O(log n). Note that the array elements will no longer be ints; instead, they'll have to be objects with a colour attribute.
I wouldn't recommend it.
Any particular reason this demands an array? You need an data structure which keeps data ordered and allows you to insert quickly. Why not a binary search tree? Or better still, a red black tree. In C++, you could use the Set structure in the Standard template library which is implemented as a red black tree. Gives you O(log(n)) insertion time and the ability to iterate over it like an array.
Intro
Consider you have a list of key/value pairs:
(0,a) (1,b) (2,c)
You have a function, that inserts a new value between two current pairs, and you need to give it a key that keeps the order:
(0,a) (0.5,z) (1,b) (2,c)
Here the new key was chosen as the average between the average of keys of the bounding pairs.
The problem is, that you list may have milions of inserts. If these inserts are all put close to each other, you may end up with keys such to 2^(-1000000), which are not easily storagable in any standard nor special number class.
The problem
How can you design a system for generating keys that:
Gives the correct result (larger/smaller than) when compared to all the rest of the keys.
Takes up only O(logn) memory (where n is the number of items in the list).
My tries
First I tried different number classes. Like fractions and even polynomium, but I could always find examples where the key size would grow linear with the number of inserts.
Then I thought about saving pointers to a number of other keys, and saving the lower/greater than relationship, but that would always require at least O(sqrt) memory and time for comparison.
Extra info: Ideally the algorithm shouldn't break when pairs are deleted from the list.
I agree with snowlord. A tree would be ideal in this case. A red-black tree would prevent things from getting unbalanced. If you really need keys, though, I'm pretty sure you can't do better than using the average of the keys on either side of the value you need to insert. That will increase your key length by 1 bit each time. What I recommend is renormalizing the keys periodically. Every x inserts, or whenever you detect keys being generated too close together, renumber everything from 1 to n.
Edit:
You don't need to compare keys if you're inserting by position instead of key. The compare function for the red-black tree would just use the order in the conceptual list, which lines up with in-order in the tree. If you're inserting in position 4 in the list, insert a node at position 4 in the tree (using in-ordering). If you're inserting after a certain node (such as "a"), it's the same. You might have to use your own implementation if whatever language/library you're using requires a key.
I don't think you can avoid getting size O(n) keys without reassigning the key during operation.
As a practical solution I would build an inverted search tree, with pointers from the children to the parents, where each pointer is marked whether it is coming from a left or right child. To compare two elements you need to find the closest common ancestor, where the path to the elements diverges.
Reassigning keys is then rebalancing of the tree, you can do that by some rotation that doesn't change the order.
If I have N arrays, what is the best(Time complexity. Space is not important) way to find the common elements. You could just find 1 element and stop.
Edit: The elements are all Numbers.
Edit: These are unsorted. Please do not sort and scan.
This is not a homework problem. Somebody asked me this question a long time ago. He was using a hash to solve the problem and asked me if I had a better way.
Create a hash index, with elements as keys, counts as values. Loop through all values and update the count in the index. Afterwards, run through the index and check which elements have count = N. Looking up an element in the index should be O(1), combined with looping through all M elements should be O(M).
If you want to keep order specific to a certain input array, loop over that array and test the element counts in the index in that order.
Some special cases:
if you know that the elements are (positive) integers with a maximum number that is not too high, you could just use a normal array as "hash" index to keep counts, where the number are just the array index.
I've assumed that in each array each number occurs only once. Adapting it for more occurrences should be easy (set the i-th bit in the count for the i-th array, or only update if the current element count == i-1).
EDIT when I answered the question, the question did not have the part of "a better way" than hashing in it.
The most direct method is to intersect the first 2 arrays and then intersecting this intersection with the remaining N-2 arrays.
If 'intersection' is not defined in the language in which you're working or you require a more specific answer (ie you need the answer to 'how do you do the intersection') then modify your question as such.
Without sorting there isn't an optimized way to do this based on the information given. (ie sorting and positioning all elements relatively to each other then iterating over the length of the arrays checking for defined elements in all the arrays at once)
The question asks is there a better way than hashing. There is no better way (i.e. better time complexity) than doing a hash as time to hash each element is typically constant. Empirical performance is also favorable particularly if the range of values is can be mapped one to one to an array maintaining counts. The time is then proportional to the number of elements across all the arrays. Sorting will not give better complexity, since this will still need to visit each element at least once, and then there is the log N for sorting each array.
Back to hashing, from a performance standpoint, you will get the best empirical performance by not processing each array fully, but processing only a block of elements from each array before proceeding onto the next array. This will take advantage of the CPU cache. It also results in fewer elements being hashed in favorable cases when common elements appear in the same regions of the array (e.g. common elements at the start of all arrays.) Worst case behaviour is no worse than hashing each array in full - merely that all elements are hashed.
I dont think approach suggested by catchmeifyoutry will work.
Let us say you have two arrays
1: {1,1,2,3,4,5}
2: {1,3,6,7}
then answer should be 1 and 3. But if we use hashtable approach, 1 will have count 3 and we will never find 1, int his situation.
Also problems becomes more complex if we have input something like this:
1: {1,1,1,2,3,4}
2: {1,1,5,6}
Here i think we should give output as 1,1. Suggested approach fails in both cases.
Solution :
read first array and put into hashtable. If we find same key again, dont increment counter. Read second array in same manner. Now in the hashtable we have common elelements which has count as 2.
But again this approach will fail in second input set which i gave earlier.
I'd first start with the degenerate case, finding common elements between 2 arrays (more on this later). From there I'll have a collection of common values which I will use as an array itself and compare it against the next array. This check would be performed N-1 times or until the "carry" array of common elements drops to size 0.
One could speed this up, I'd imagine, by divide-and-conquer, splitting the N arrays into the end nodes of a tree. The next level up the tree is N/2 common element arrays, and so forth and so on until you have an array at the top that is either filled or not. In either case, you'd have your answer.
Without sorting and scanning the best operational speed you'll get for comparing 2 arrays for common elements is O(N2).
So, suppose you have a collection of items. Each item has an identifier which can be represented using a bitfield. As a simple example, suppose your collection is:
0110, 0111, 1001, 1011, 1110, 1111
So, you then want to implement a function, Remove(bool bitval, int position). For example, a call to Remove(0, 2) would remove all items where index 2(i.e. 3rd bit) was 0. In this case, that would be 1001, only. Remove(1,1) would remove 1110, 1111, 0111, and 0110. It is trivial to come up with an O(n) collection where this is possible (just use a linked list), with n being the number of items in the collection. In general the number of items to be removed is going to be O(n) (assuming a given bit has a ≥ c% chance of being 1 and a ≥ c% chance of being 0, where c is some constant > 0), so "better" algorithms which somehow are O(l), with l being the number of items being removed, are unexciting.
Is it possible to define a data structure where the average (or better yet, worst case) removal time is better than O(n)? A binary tree can do pretty well (just remove all left/right branches at the height m, where m is the index being tested), but I'm wondering if there is any way to do better (and quite honestly, I'm not sure how to removing all left or right branches at a particular height in an efficient manner). Alternatively, is there a proof that doing better is not possible?
Edit: I'm not sure exactly what I'm expecting in terms of efficiency (sorry Arno), but a basic explanation of it's possible application is thus: Suppose we are working with a binary decision tree. Such a tree could be used for a game tree or a puzzle solver or whatever. Further suppose the tree is small enough that we can fit all of the leaf nodes into memory. Each such node is basically just a bitfield listing all of the decisions. Now, if we want to prune arbitrary decisions from this tree, one method would be to just jump to the height where a particular decision is made and prune the left or right side of every node (left meaning one decision, right meaning the other). Normally in a decision tree you only want to prune subtree at a time (since the parent of that subtree is different from the parent of other subtrees and thus the decision which should be pruned in one subtree should not be pruned from others), but in some types of situations this may not be the case. Further, you normally only want to prune everything below a particular node, but in this case you'll be leaving some stuff below the node but also pruning below other nodes in the tree.
Anyhow, this is somewhat of a question based on curiousity; I'm not sure it's practical to use any results, but am interested in what people have to say.
Edit:
Thinking about it further, I think the tree method is actually O(n / logn), assuming it's reasonably dense. Proof:
Suppose you have a binary tree with n items. It's height is log(n). Removing half the bottom will require n/2 removals. Removing the half the row above will require n/4. The sum of operations for each row is n-1. So the average number of removals is n-1 / log(n).
Provided the length of your bitfields is limited, the following may work:
First, represent the bitfields that are in the set as an array of booleans, so in your case (4 bit bitfields), new bool[16];
Transform this array of booleans into a bitfield itself, so a 16-bit bitfield in this case, where each bit represents whether the bitfield corresponding to its index is included
Then operations become:
Remove(0, 0) = and with bitmask 1010101010101010
Remove(1, 0) = and with bitmask 0101010101010101
Remove(0, 2) = and with bitmask 1111000011110000
Note that more complicated 'add/remove' operations could then also be added as O(1) bit-logic.
The only down-side is that extra work is needed to interpret the resulting 16-bit bitfield back into a set of values, but with lookup arrays that might not turn out too bad either.
Addendum:
Additional down-sides:
Once the size of an integer is exceeded, every added bit to the original bit-fields will double the storage space. However, this is not much worse than a typical scenario using another collection where you have to store on average half the possible bitmask values (provided the typical scenario doesn't store far less remaining values).
Once the size of an integer is exceeded, every added bit also doubles the number of 'and' operations needed to implement the logic.
So basically, I'd say if your original bitfields are not much larger than a byte, you are likely better off with this encoding, beyond that you're probably better off with the original strategy.
Further addendum:
If you only ever execute Remove operations, which over time thins out the set state-space further and further, you may be able to stretch this approach a bit further (no pun intended) by making a more clever abstraction that somehow only keeps track of the int values that are non-zero. Detecting zero values may not be as expensive as it sounds either if the JIT knows what it's doing, because a CPU 'and' operation typically sets the 'zero' flag if the result is zero.
As with all performance optimizations, this one'd need some measurement to determine if it is worthwile.
If each decision bit and position are listed as objects, {bit value, k-th position}, you would end up with an array of length 2*k. If you link to each of these array positions from your item, represented as a linked list (which are of length k), using a pointer to the {bit, position} object as the node value, you can "invalidate" a bunch of items by simply deleting the {bit, position} object. This would require you, upon searching the list of items, to find "complete" items (it makes search REALLY slow?).
So something like:
[{0,0}, {1,0}, {0,1}, {1, 1}, {0,2}, {1, 2}, {0,3}, {1,3}]
and linked from "0100", represented as: {0->3->4->6}
You wouldn't know which items were invalid until you tried to find them (so it doesn't really limit your search space, which is what you're after).
Oh well, I tried.
Sure, it is possible (even if this is "cheating"). Just keep a stack of Remove objects:
struct Remove {
bool set;
int index;
}
The remove function just pushes an object on the stack. Viola, O(1).
If you wanted to get fancy, your stack couldn't exceed (number of bits) without containing duplicate or impossible scenarios.
The rest of the collection has to apply the logic whenever things are withdrawn or iterated over.
Two ways to do insert into the collection:
Apply the Remove rules upon insert, to clear out the stack, making in O(n). Gotta pay somewhere.
Each bitfield has to store it's index in the remove stack, to know what rules apply to it. Then, the stack size limit above wouldn't matter
If you use an array to store your binary tree, you can quickly index any element (the children of the node at index n are at index (n+1)*2 and (n+1)*2-1. All the nodes at a given level are stored sequentially. The first node at at level x is 2^x-1 and there are 2^x elements at that level.
Unfortunately, I don't think this really gets you much of anywhere from a complexity standpoint. Removing all the left nodes at a level is O(n/2) worst case, which is of course O(n). Of course the actual work depends on which bit you are checking, so the average may be somewhat better. This also requires O(2^n) memory which is much worse than the linked list and not practical at all.
I think what this problem is really asking is for a way to efficiently partition a set of sets into two sets. Using a bitset to describe the set gives you a fast check for membership, but doesn't seem to lend itself to making the problem any easier.