One of the most important data mining problems is association rules. In this paper we consider discovering rules from large transaction databases. The problem can be decomposed into two sub-problems: find itemsets and generate itemsets. second sub-problem easier one complexity determined by paper, suggest Apriori-based algorithm for Actually, a new procedure generation which more efficient than appropriate original Apriori algorithm. For its implementation, modified sort-merge-join...

The aim of this study is to establish the scientific benefit royal jelly (RJ) on prostatic-specific antigen (PSA), post-void residual (PVR) volume and International Prostate Symptom Score (IPSS) in benign prostatic hyperplasia. For study, a group 40 men were administered 38 mg RJ over period three months, their PSA values, prostate volumes transitory zones, PVR IPPS values measured at end first month, third month. results confirm potential reducing scores improving IPSS values. Since use did...

Prostate cancer (PCa) is the second most commonly diagnosed among men worldwide. Despite its high incidence rate, molecular basis of PCa onset and progression remains little understood. Genome-wide association studies (GWAS) have greatly contributed to identification single nucleotide polymorphisms (SNP) associated with risk. Several GWAS identified 8q24 as one significant PCa-associated regions. The aim this study was evaluate SNP rs378854 at risk in Serbian population. population included...

Results of recent studies confirmed that oxidative stress negatively affects sperm motility and causes DNA damage. Produced by nitric oxide synthase 3 (NOS3), is considered to be one the important mediators in testis tissue. The aim this study was assess possible association three genetic variants (rs2070744, rs1799983 intron variant 4a/4b) NOS3 gene infertility occurrence two groups infertile men (idiopathic azoospermia oligoasthenozoospermia) fertile controls. Genotypes for...

Layout analysis is essential step in information extraction from scanned document images. In this paper we propose an algorithm for aligning layouts generated with different OCR engines. The main requirement to always generate the same layout given image regardless of engine used processing. that way documents, heavily dependent on fields positions document, does not depend specific engine. other words, it sufficient maintain universal extractor knowledge and necessary train explicitly...

We aimed to analyze the correlation of perineural invasion on transrectal ultrasound guided prostate biopsy with predictors biochemical cancer recurrence, as well its impact clinical outcomes, for non-metastatic cancer. For study, patients (N = 86) were recruited into group I and underwent open retropubic prostatectomy, regardless stage; cases but without biopsy, who received radical prostatectomy treatment modality, placed II (n 90). Perineural was detected preoperatively in 43% that...

In this paper we suggest a new method for frequent itemsets mining, which is more efficient than well known Apriori algorithm. The based on special structure called Rymon tree. For its implementation, modified sort-merge-join Finally, explain how support measure, used in algorithm, gives statistically significant itemsets.

In this paper we present novel mathematical model of the frequent itemset mining problem. The is based on set intersection terminology and theory. We use linear algebra method in order to prove our results.

In this study, we present results of applying data mining techniques on meteorological dataset obtained from the Institute Hydrometeorology and Seismology Montenegro.The covers measurements taken all 11 main stations in Montenegro for period 2010-2015.We build new climate classification system based decision tree.The is simpler (i.e.uses fewer attributes) more accurate than well-known Köppen system.In addition, propose a novel procedure network construction.Finally, identify regions within...

The formulation of the 3 n -1 problem is simple but no one has found solution yet.This paper transforms original into its equivalent so that it becomes more suitable for computer validation.A new algorithm proposed and implemented.The hypothesis tested proven to be valid numbers -1, conclusive with number 32768 -1.

In this paper we present efficient algorithm for enumerating frequent pairs in database of transactions. Counting support itemset can be dominant step mining occupying some cases greater that 90% total execution time. For implementation use one-dimensional array and define one-to-one function maps 2-itemsets to indexes preserving lexicographic order. way are not generated explicitly; they represented by while their counts stored appropriate elements.

Frequent itemset mining has been a focused theme in data research for years. It was first proposed market basket analysis the form of association rule mining. Since proposal this new task and its associated efficient algorithms, there have hundreds followup publications. In paper we further develop ideas presented [1]. [1] consider two problems from linear algebra, namely set intersection problem scalar product make comparisons to frequent task. formulate prove theorems that estimate number...

In this paper we present the first case study on collaboration network of researchers at University Montenegro - UoM. We identify largest clusters or groups scientists that are interested in same topic, using Girvan-Newman algorithm. The results show these constantly grow over period 2005-2015 and moment they occupy more than 50% authors from It indicates there exists increasingly between UoM network. But, according to classification real networks is "sub-critical" regime, still far away...

In this paper we present an approach for novelty detection in text data. The can also be considered as semi-supervised anomaly because it operates with the training dataset containing labelled instances known classes only. During phase classification model is learned. It assumed that at least two exist available dataset. testing are classified normal or anomalous based on classifier confidence. other words, if cannot assign any of class labels to given instance sufficiently high confidence...

In this paper the problem of measuring factor importance on patient length stay in an emergency department is discussed. Historical dataset contains average per day. Factors are agreed with domain expert. The task to provide factors? impact measure specific day that does not belong historical (new observation) and for higher than specified threshold. Observations represented as multidimensional numeric vectors. Each dimension represents factor. basic idea consists identifying appropriate...

In this paper we establish novel theoretical framework for frequent itemset mining problem, as well new and efficient join-based algorithm. The algorithm maintains one-dimensional array rank, starting from k=2nd iteration. uses combinatorial number schema to map k-candidates indexes of array. At the index r array, stores RANKC rth candidate in lexicographic order, so , where is order iteration k. Having hand rank will join two candidates iff their ranks are equal, making an operation. Also,...