Data mining for text categorization with semi-supervised agglomerative hierarchical clustering

This paper presents semi-supervised agglomerative hierarchical clustering algorithm using clusterwise tolerance based pairwise constraints. In semi-supervised clustering, pairwise constraints, that is, must-link and cannot-link, are frequently used in order to improve clustering properties. From that sense, we will propose another way named clusterwise tolerance based pairwise constraints to handle must-link and cannot-link constraints inL2-space. In addition, we will propose semi-supervised agglomerative hierarchical clustering algorithm based on it. We will, moreover, show the effectiveness of the proposed method through numerical examples.

Regionalization of watersheds by hybrid-cluster analysis

This work aims to expand the knowledge of the area of data analysis through persistence homology and representations of directed graphs. To be specific, we looked for how we can analyze homology cluster groups using agglomerative Hierarchical Clustering algorithms and methods. Additionally, the Wine data, which is offered in R studio, was analyzed using various cluster algorithms such as Hierarchical Clustering, K-Means Clustering, and PAM Clustering. The goal of the analysis was to find out which cluster's method is proper for a given numerical dataset. We tried to find the agglomerative hierarchical clustering method by testing the data that will be the optimal clustering algorithm among these three; K-Means, PAM, and Random Forest methods. By comparing each model's accuracy value with cultivar coefficients, we concluded that K-Means methods are the most helpful when working with numerical variables. On the other hand, PAM clustering and Gower with Random Forest are the most beneficial approaches when using categorical variables. These tests can determine the optimal number of clustering groups, given the data set, and by doing the proper analysis. Using those the project, we can apply our method to several industrial areas such that clinical, business, and others. For example, people can make different groups based on each patient who has a common disease, required therapy, and other things in the clinical society. Additionally, people can expect to get several clustered groups based on the marginal profit, marginal cost, or other economic indicators for the business area.

Reducing package‐related cost is essential for various companies and institutions. Different packages are usually designed separately for each and every product, which results in less cost‐effective packaging systems. In this study, a data mining model with three clustering algorithms was developed to modularize a packaging system by reducing the variety of packaging sizes. The three algorithms were k‐means clustering, agglomerative hierarchical clustering and self‐organizing feature map. The package models with similar shapes and sizes were clustered automatically and replaced by one package model with a size that suited them all. The study also analysed the financial effects including the purchasing and inventory costs of the package material and the transportation cost of the packaged products. The case study was carried out at Ericsson to select the best clustering algorithm of the three and to test the effectiveness and applicability of the proposed model. The results show that the packaging system modularized by the agglomerative hierarchical clustering algorithm is more cost‐effective in this case compared with the ones modularized by the other two clustering algorithms and with the one without modularization. Copyright © 2017 John Wiley & Sons, Ltd.

Classifying an unknown object in image retrieval systems using the nearest neighbour classifier would be very time consuming when the number of the objects within the associated database is high. Generating a dendrogram using a Hierarchical Agglomerative Clustering (HAC) algorithm and searching the database images from coarse to fine resolutions using image pyramids are two important groups of techniques widely used for dealing with this problem. In this paper, a novel algorithm is proposed by combining these methods within the framework of a face recognition system. The search process is performed in a coarse-to-fine manner using image pyramids. On the bottom level of the pyramid (the finest resolution), a set of dendrograms is formed using the HAC algorithm. Our experimental studies show that the recognition process can be speeded up by a factor of around 65 compare to the basic nearest neighbour classifier. In such a condition, however, the recognition rate is slightly reduced.

The study analysed salesmen itinerary of a confectionery supplier in Indonesia. Channel management and robust product distribution are important elements for confectionery business, since growth heavily depends on healthy product distribution. To support product distribution, optimising salesmen itinerary plays an important role in the firm's sales operation. The study was limited to optimising salesmen schedule for its wholesalers' outlet type, since it contributed to 78% of its revenue. The analyses were carried out using 'cluster first-route second' approach. For clustering the customer nodes, it recommended agglomerative hierarchical clustering (AHC) method. For routing within each customer's cluster, it tested two algorithms, i.e., Clarke-Wright and weighted-serving time-visual approach. The result showed that AHC-weighted-serving-time-visual algorithm was as good as AHC-Clarke-Wright with 2.1% cheaper variable cost. The proposed algorithm was adapted from practical insight and easier to be understood.

Spoken dialogue systems, which provide automated customer services at call centers, have become more prevalent. It is time consuming to determine a set of call types for the dialogue system by analyzing a large volume of unstructured spoken utterances. Traditional hierarchical agglomerative clustering (HAC) algorithms can bootstrap the call types in an unsupervised way, yet the time and space complexities are huge, especially for a large data set. Based on our observation that spoken utterances containing less than ten terms are common in the spoken dialogue system, we propose an efficient HAC algorithm for short utterances. By utilizing the particular properties of short utterances, we significantly reduce both the time and the space complexities of the clustering algorithm.

The authors show that most hierarchical agglomerative clustering (HAC) algorithms follow a 90-10 rule where roughly 90% iterations from the beginning merge cluster pairs with dissimilarity less than 10% of the maximum dissimilarity. We propose two algorithms: 2-phase and nested, based on partially overlapping partitioning (POP). To handle high-dimensional data efficiently, we propose a tree structure particularly suitable for POP. Extensive experiments show that the proposed algorithms reduce the time and memory requirement of existing HAC algorithms significantly without compromising accuracy.

The proliferation of big data and the growth of e-commerce have intensified the challenges associated with extracting actionable data for personalised recommendations and decision-making. With data-driven marketing strategies, understanding and predicting customer behaviour has become paramount for maintaining competitive advantage. This study leverages business analytics tools, focusing on Recency, Frequency, and Monetary (RFM) Analysis, alongside K-Means and Hierarchical (Agglomerative) Clustering algorithms, to segment customer transactional data. Data normalisation, a critical step for accurate clustering, was performed using log transformation and the Power Transformer technique with the Yeo-Johnson parameter, the latter proving more effective for handling both positively and negatively skewed data, enhancing data normalisation and suitability for analysis. This study reveals that RFM Analysis with Hierarchical Clustering outperforms K-Means Clustering, achieving a Silhouette Score of 0.47 and a Calinski–Harabasz Index of 3787.1, indicating a more accurate identification of customer segments. RFM Analysis alone generated eight clusters, while integrating RFM Analysis with both Hierarchical Clustering and K-Means generated three similar-sized clusters with interchanged labels. These metrics highlight the proficiency of Hierarchical Clustering in identifying unique customer segments and customising marketing strategies. The findings indicate that the RFM-Hierarchical Clustering approach enhances segmentation precision and facilitates more refined and effective marketing strategies.

Hybrid precoding design can be challenging for broadband millimeter-wave (mmWave) massive MIMO due to the frequency-flat analog precoder in radio frequency (RF). Prior broadband hybrid precoding work usually focuses on fully-connected array (FCA), while seldom considers the energy-efficient partially-connected subarray (PCS) including the fixed subarray (FS) and dynamic subarray (DS). Against this background, this paper proposes a machine learning based broadband hybrid precoding for mmWave massive MIMO with DS. Specifically, we first propose an optimal hybrid precoder based on principal component analysis (PCA) for the FS, whereby the frequency-flat RF precoder for each subarray is extracted from the principle component of the optimal frequency-selective precoders for fully-digital MIMO. Moreover, we extend the PCA-based hybrid precoding to DS, where a shared agglomerative hierarchical clustering (AHC) algorithm developed from machine learning is proposed to group the DS for improved spectral efficiency (SE). Finally, we investigate the energy efficiency (EE) of the proposed scheme for both passive and active antennas. Simulations have confirmed that the proposed scheme outperforms conventional schemes in both SE and EE.

To solve name ambiguity problems and improve the performance of person name disambiguation, we propose a three-stage clustering algorithm in the paper. In the first stage, organizations and locations (OLs) are used to cluster documents about the same person, therefore some texts with more resemblance will be assigned to one category. This stage is simply document clustering based on the similarity of OLs. In the second stage, the clustered documents are used as a new data source from which some novel features (like co-author names) are extracted. We used these new extracted features to make additional clustering between documents. Meanwhile, a method was proposed to solve name ambiguity problems by using social networks construction based on the relationships among co-authors. In the third stage, Web pages are further clustered using content-based hierarchical agglomerative clustering (HAC) algorithm, then analyzing the useful content including title and abstract and keywords (TAKs) to disambiguate the ambiguous names. Experimental results show that our three-stage clustering algorithm can availably enhance the performance of person name disambiguation.

Customer segmentation plays an important role in customer relationship management. It allows companies to design and establish different strategies to maximize the value of customers. Customer segmentation refers to grouping customers into different categories based on shared characteristics such as age, location, spending habit and so on. Similarly, clustering means putting things together in such a way that similar type of things remain in the same group. In this study, a machine learning (ML) hierarchical agglomerative clustering (HAC) algorithm is implemented in the R programming language to perform customer segmentation on credit card data sets to determine the appropriate marketing strategies.

The automatic or semi-automatic construction of ontology has become a research topic of interest in recent years. This paper describes a mechanism for constructing domain specific ontologies automatically based on web data and generic ontology. Firstly, we employ the hierarchical agglomerative clustering(HAC) algorithm, clustering web pages hierarchically and resulting in a binary tree. Then an algorithm is proposed, which selectd from the binary tree the significative nodes as topics implying concepts of domain interests. Lastly, the Chinese generic ontology, HowNet, is introduced to evolve the topics (together with their hierarchical structures) into domain ontology. We experiment our method in the field of computer hardware based on web pages collected from Chinese BtoC web sites. An in-depth discussion on the experiment results is also given.

Both Earth and Mars host populations of co-orbital minor bodies. A large number of present-day Mars co-orbitals is probably associated with the fission of the parent body of Mars Trojan 5261 Eureka (1990 MB) during a rotation-induced Yarkovsky–O’Keefe–Radzievskii–Paddack (YORP) break-up event. Here, we use the statistical distributions of the Tisserand parameter and the relative mean longitude of Mars co-orbitals with eccentricity below 0.2 to estimate the importance of rotation-induced YORP break-up events in Martian co-orbital space. Machine-learning techniques (k-means++ and agglomerative hierarchical clustering algorithms) are applied to assess our findings. Our statistical analysis identified three new Mars Trojans: 2009 SE, 2018 EC4, and 2018 FC4. Two of them, 2018 EC4 and 2018 FC4, are probably linked to Eureka but we argue that 2009 SE may have been captured, so it is not related to Eureka. We also suggest that 2020 VT1, a recent discovery, is a transient Martian co-orbital of the horseshoe type. When applied to Earth co-orbital candidates with eccentricity below 0.2, our approach led us to identify some clustering, perhaps linked to fission events. The cluster with most members could be associated with Earth quasi-satellite 469219 Kamo‘oalewa (2016 HO3) that is a fast rotator. Our statistical analysis identified two new Earth co-orbitals: 2020 PN1, which follows a horseshoe path, and 2020 PP1, a quasi-satellite that is dynamically similar to Kamo‘oalewa. For both Mars and Earth co-orbitals, we found pairs of objects whose values of the Tisserand parameter differ by very small amounts, perhaps hinting at recent disruption events. Clustering algorithms and numerical simulations both suggest that 2020 KZ2 and Kamo‘oalewa could be related.

Although semi-supervised classification has widely been studied by many researchers, semi-supervised agglomerative hierarchical clustering is not popular. Two methods to introduce pairwise constraint to agglomerative hierarchical clustering have been proposed so far. The first method is to modify distance between two objects that should be in difference clusters by using a kernel function. The second method is to add a penalty term to similarity measure. In addition, difference of linkage methods of agglomerative hierarchical clustering should be compared to observe how the difference affects the resulting clusters. In this paper, we compare different linkage methods with the above two methods for pairwise constraints. Moreover asymmetric similarity measure is considered. Effects of pairwise constraints are shown by simple examples.