ai learning algorithms

algorithm to initially estimate the number of K.2) How to choose the initial K pointsThe common algorithm is random selection. But often the effect is not very good, also can be similar to the method, the line uses the hierarchical clustering algorithm to divide the K clusters, and uses these clusters ' centroid as the initial centroid.3) method of calculating distancesCommonly used such as European distance, cosine angle similarity degree.4) Algorithm Stop conditionThe maximum number of iterat

other.Suppose we choose the attribute R as the split attribute, DataSet D, R has K different values {v1,v2,..., Vk}, so d according to the value of R into K-group {d1,d2,..., Dk}, after splitting by R, the amount of information required to separate the different classes of DataSet D is:information gain is defined as before and after the split, two of the amount is only poor:The following example uses Python to illustrate a decision tree construct using the information gain method:The main steps

[],intN) {/ * Merge sort * /ElementType *tmpa; Tmpa = (ElementType *)mallocNsizeof(ElementType));if(Tmpa! = NULL) {Msort (A, Tmpa,0N1); Free(TMPA); }Else printf("Not Enough space");}intMain () {ElementType a[100001], N; clock_t start, finish;Cin>> N; for(inti =0; i Cin>> A[i]; start = Clock (); Insertionsort (A, N); finish = Clock ();DoubleDuration = (finish-start)/clocks_per_sec;//cout cout0]; for(inti =1; i cout" "//cout }VI. TestingEach test point data is as follows to compare the perf

Sorting algorithmafter mastering the functions of the max_heapify () function and the Build_max_heap () function, we have the basis of mastering the heap sorting algorithm. heap Sorting algorithm process: Using Build_max_heap () to build the largest heap of A[,1......N], where n=a.length; Exchange A[1] and a[length] values so that the largest element is in the last position of the heap; Heap_size = heap_size-1; Reduce the number of heap elements by 1, that is, exclude the maxim

the register to compare, meet the preceding paragraph greater than latter, the first 7 position with 6 cover off;The sequence is: 3,5,6,6,7,8Then 5 and the register to compare, meet the preceding paragraph greater than latter, the first 6 position with 5 cover off;The sequence is: 3,5,5,6,7,8Then 3 and the register to compare, meet the preceding paragraph greater than latter, the first 5 position with 3 cover off;The sequence is: 3,3,5,6,7,8Finally, the position of the first 3 is overwritten wi

; Print (A, length);}voidMain () {intarr[7] = {1,5,6, -, the,122, the }; intb[7]; Print (arr,7); ////write a bubble first //Bubble (arr, 7); ////Block Row //Quick (arr, 0,6); //cout ////Merge //MergeSort (arr, B, 7); ////Select //Selection (arr, 7); //before the heap is sorted, first recognize what is the heap and then build the heap,//After the heap is completed, the heap can be sortedHeapsort (arr,7); Print (arr,7); intA =sizeof(arr); CIN>>A;}Reference:http://blog.csdn.ne

does not support the comparison size.-priority the priority counter, ensuring that the smallest element function in the queue (the most important) is always placed in _queue[0]. heapq.heappush()and the first element is heapq.heappop() inserted and deleted separately on the queue, the _queue first element of the _queue is always the smallest, ensuring that the queue _queue the first element with the highest priority.Reference: "Python Cookbook" 3rd EditionHttp://python3-cookbook.readthedocs.io/z

--found 3 meet the requirements, then i++ found I=j, indicating that this round of movement is over. Now the base number 6 and 3 are exchanged, the result: 3 1 2 5 4 6 9 7 ten 8; now 6 The left is yet.　　　　Figure: Adam and Eve finally have a intersection(6) We take the number of the left of 6 out first: 3 1 2 5 4, this time with 3 as the base number to adjust, so that 3 left the number of 3 5 4(7) Then 2 1 key out to reorganize, the results obtained: 1 2(8) The remaining sequence on the right:

] = 1 initialized to 1@DATE:2017/9/2*/#includeusing namespacestd;intarr[10010],tmp[10010];intLen;/*status definition: Used with the TMP array Tmp[i]: For all lis of length I, his results are the smallest possible, if the smaller the more easily be taken into the TMP element is strictly incremented State transfer: if (dp[j] = i) Tmp[i] = min (arr[j]) results Lookup--using a two-point method to find if (Tmp[i] *///binary lookup, binary lookup in tmp arr[i]//updating the TMP arrayvoidBisearch (intx

state, what would happen if we took the strips [3,5] off? We found that the final part of the shadow was still 6. 。。。 Here is the code for the delete action:1 voidDelintLintRintStep) {2 if(L==a[step].leftr==a[step].right) {//fully cover3a[step].cover--;return;4 }5 if(A[step].left==a[step].right)return;6 intMid= (a[step].left+a[step].right)/2;7 if(R2);8 Else if(l>=mid+1) del (l,r,step*2+1);9 Else{TenDel (l,mid,step*2); OneDel (mid+1, r,step*2+1); A } -}Assuming th

application thread exists in the contents of the set logs, and modify the corresponding remembered sets, this step needs to pause the application, parallel running.Survival Object calculation and cleanup ( Live Data counting and Cleanup )It should be noted that in G1, it is not that final marking pause is executed, it is certain to perform cleanup this step, because this step needs to suspend the application, G1 in order to achieve quasi-real-time requirements, It is necessary to reasonably pla

Logistic regression is used to classify, and linear regression is used to return.Linear regression is the addition of the properties of the sample to the front plus the coefficients. The cost function is the sum of squared errors. Therefore, in the minimization of the cost function, you can directly derivative, so that the derivative equals 0, as follows:Gradient descent can also be used to learn the same gradient as the logistic regression form.Advantages of linear regression: simple calculatio

Platform: Ubuntu 12.04/GCC 4.6.7 Insert Sort1#include 2#include 3#include 4 using namespacestd;5 6Template 7 voidInsertsort (vectorVEC) {8 //vector9 for(Auto j=1; J)Ten { OneT key=Vec[j]; A inti=j-1; - while(i>=0vec[i]>key) - { thevec[i+1]=Vec[i]; -i--; - } -vec[i+1]=key; + } - } + ATemplate at voidPrintvector (ConstvectorVEC) - { - for(Auto i=0; I) -cout" "; - -coutEndl; in } - to intMain () + { - the //void Insertsort (vect

], and is the going to enter the i th check. And we can assume that we have + already get the maximum subarray of the a[1...i-1].supposing the A maxisum Subarray of A[1..i-1] is a[p. I-1]. the + we have aleady get procedure are going to the ith check - the solution of a[1..i], and we'll move to check a[i]. $ relationship between Sum[i-1] and Sum[i]? From our experience, $ - * If sum[i-1] - A[p...i-1,i] must is smaller than that of Subarray a[p. I-1]. So we the Modify t

1.河内之塔..2.Algorithm Gossip: 费式数列.3. 巴斯卡三角形4.Algorithm Gossip: 三色棋5.Algorithm Gossip: 老鼠走迷官（一）6.Algorithm Gossip: 老鼠走迷官（二）7.Algorithm Gossip: 骑士走棋盘8.Algorithm Gossip: 八皇后9.Algorithm Gossip: 八枚银币.10.Algorithm Gossip: 生命游戏.11.Algorithm Gossip: 字串核对12.Algorithm Gossip: 双色、三色河内塔13.Algorithm Gossip: 背包问题（Knapsack Problem14.Algorithm Gossip: 蒙地卡罗法求 PI15.Algorithm Gossip: Eratosthenes 筛选求质数16.Algorithm Gossip: 超长整数运算（大数运算）.17.Algorithm Gossip: 长 PI.18.Algorithm Gossip: 最大公因数、最小公倍数、因式分解19.Algorithm Gossi

Any sort of comparison algorithm, the worst-case operating time limit is Ω (NLGN)Count sortAssuming that each of the N input elements is an integer between 0 and K, K is an integer, and when K=o (N), the run time of the count sort is θ (n)1 //input Array A[1..N], holding the sorted result array B[1..N], temporary storage C[0..K]2counting-SORT (a,b,k)3 fori←0to K4 Doc[i]←05 forj←1To Length[a]6 Doc[a[j]]←c[a[j]]+17 fori←1to K8 Doc[i]←c[i]+c[i-1]9 forJ←length[a] Downto1Ten D

Quick Sort1 QUICKSORT (a,p,r)2 ifpR3 Then q←partition (a,p,r)4QUICKSORT (a,p,q-1)5QUICKSORT (a,q+1, R)6 7 PARTITION (a,p,r)8 X←a[r]9i←p-1Ten forJ←p to R1 One Do ifa[j]x AThen i←i+1 - Exchange A[i]↔a[j] -Exchange a[i+1]↔a[r] the returni+1Randomize The fast sort, no longer chooses R as the main element, but randomly chooses a main element and puts it in R position1randomized-PARTITION (a,p,r)2 i←random (p,r)3 Exchange A[r]↔a[i]4 returnPARTITION (a,p,r)5 6randomized-QUICKSORT (a,p,r)7 ifpR8Th

]] - Thenω←right[p[x]] - ifcolor[ω]=RED -Then color[ω]=BLACK - color[p[x]]←red inleft-ROTATE (t,p[x]) - Ω←right[p[x]] to ifColor[left[ω]]=black and color[right[ω]]=BLACK + Then color[ω]←red - X←p[x] the Else ifcolor[right[ω]]=BLACK * Then color[left[ω]]←black $ color[ω]←redPanax Notoginsengright-ROTATE (t,ω) - Ω←right[p[x]] the Color[ω]←color[p[x]] + Color[p[x]]←black A Color[right[ω]]←black theleft-ROTATE (t,p[x]) + X←root[t] - Else(Same asThen cl

);Panax Notoginseng Else if(l>=mid+1) Ret=query_min (l,r,rt*2+1); - ElseRet=min (Query_min (l,mid,rt*2), Query_min (mid+1, r,rt*2+1)); the returnret; + } A } the intMain () { + intn,m,q,p,v; - while(SCANF ("%d", n)! =EOF) { $Buildtree (1N1); $scanf"%d",m); - for(intI=1; i){ -scanf"%d%d",p,v); theInsert (P,v,1); - }Wuyiscanf"%d",q); the while(q--){ - intb; Wuscanf"%d%d",a,b); -printf"%d\n", Query_min (A, B,1)); About

result.So what do we do with the program language output?int sum =0,n = 100;for (int i=1;i{Sum =sum+i;}coutGaussian algorithmint i,sum=0,n=100;sum = (1+n) *N/2;cout Execute only onceWhat is an algorithm?A description of the solution steps for a specific problem, represented as a finite sequence of instructions in the computer, and each instruction represents one or more operationsFive basic featuresInput, output, poor, deterministic and feasibleInputAlgorithm Local area 0 or more inputsvoid Tes