OPTICS on Sequential Data: Experiments and Test Results

Abstract

The Web has enormous, various and knowledgeable data for data mining research. Clustering web usage data is useful to discover interesting patterns pertaining to user traversals, behaviour and their usage characteristics. Moreover, users accesses web pages in an order in which they are interested and hence incorporating sequence nature of their usage is crucial for clustering web transactions. In this paper we present OPTICS ("Ordering Points To Identify the Clustering Structure") algorithm to find density based clusters on a web usage data on MSNBC.COM website which is a free news data website with so different categories of news).The clusters are generated by OPTICS algorithm . The average of inter cluster and intra cluster are Calculated. the results are compared with different similarity measures like Euclidean , Jaccard, projected Euclidean, cosine and fuzzy similarity Finally showed behavior of clusters that made by OPTICS algorithm on a sequential data in a web usage domain. we performed a variety of experiments in the context of density based clustering , quantify our results by the way of explanation s and list conclusions. KeywordsClustering algorithm OPTICS, Ordering Points To Identify the Clustering Structure, Sequence mining. Average Inter cluster, Intra cluster. INTRODUCTIONThe web is a huge database for research about relationship between objects, people, socials, companies, relations, marketing, management, knowledge and etc. Clustering is a one of the ways to collecting subsets of data that have some common attributes and find hidden patterns to create knowledge from databases. Different types of data clustering are: Hierarchical, Partitional, Density-based, Sub-space clustering and etc. In this paper we use a density-based clustering algorithm that name is OPTICS algorithm to clustering a dataset from msnbc.com website. Because this algorithm is a density based algorithm and our data is a density based data. First of all we downloaded data (around 40'000 records) from the msnbc.com website and then created a dataset file. We did some preprocessing on the dataset to extract impossible data combinations for example it is possible that some attributes have never been used by the users. After accredit from the data we apply OPTICS algorithm on the new dataset to cluster the data. Finally use Euclidean distance measure, projected Euclidean distance, cosine similarity and Fuzzy dissimilarity to compare the results for intra cluster and inter cluster analyze and visualize data and graphs. RELATED WORK OPTICSThe OPTICS algorithm (Ordering Points To Identify the Clustering Structure) algorithm designed by Mihael Ankerst, Markus M. Breunig, Hans-Peter Kriegel and Jörg Sander. Its basic idea is similar to DBSCAN (Density-Based Spatial Clustering of Applications with Noise) to specify clusters and noises of a special database. OPTICS defines a cluster with the base of density. In this algorithm point "p" is a cluster if it contains minimum points (call "MinPts") that are not farther than the defin...

Prediction of users webpage access behaviour using association rule mining

“…Their results revealed that each of the clusters contained observations with specific common characteristics and improved the algorithm efficiency by identifying the initial cluster centers. Santhisree & Damodaram (2010) presented an algorithm named OPTICS ("Ordering Points To Identify the Clustering Structure") to find density based clusters on the web usage data from the MSNBC.COM website. The average of inter cluster and intra cluster distance was calculated and the results were then compared with different similarity measures like Euclidean, Jaccard, projected Euclidean, cosine and fuzzy similarity to find the similarity between clusters and the results were visualized graphically to predict the user behaviour.…”

Application of Modified OPTICS Algorithm in E-Commerce Sites Classification and Evaluation

“…The OPTICS is a density-based algorithm that starts at a randomly selected node and expands towards the most data-intensive region, eventually organizes all nodes into an ordered sequence, and then clusters different nodes (Ankerst et al, 1999).Thus, the algorithm is widely used in many fields. K. Santhisree (2010) applied OPTICS to the text data clustering. WangPin, HuangYan (2011) use the improved algorithm OPTICS clustering MPSK/MAPSK signal constellation.…”