Despite the prevalent assumption of uniform variable importance in long-term time series forecasting models, real-world applications often exhibit asymmetric causal relationships and varying data acquisition costs. Specifically, cost‐effective exogenous data (e.g., local weather) can unilaterally influence dynamics of endogenous variables, such as lake surface temperature. Exploiting these links enables more effective forecasts when exogenous inputs are readily available. Transformer-based models capture long-range dependencies but incur high computation and suffer from permutation invariance. Patch-based variants improve efficiency yet can miss local temporal patterns. To efficiently exploit informative signals across both the temporal dimension and relevant exogenous variables, this study proposes XLinear, a lightweight time series forecasting model built upon Multi-Layer Perceptrons (MLPs). XLinear uses a global token derived from an endogenous variable as a pivotal hub for interacting with exogenous variables, and employs MLPs with sigmoid activation to extract both temporal patterns and variate-wise dependencies. Its prediction head then integrates these signals to forecast the endogenous series. We evaluate XLinear on seven standard benchmarks and five real-world datasets with exogenous inputs. Compared with state-of-the-art models, XLinear delivers superior accuracy and efficiency for both multivariate forecasts and univariate forecasts influenced by exogenous inputs.

Recent semi-supervised anomaly detection methods that are trained using small labeled anomaly examples and large unlabeled data (mostly normal data) have shown largely improved performance over unsupervised methods. However, these methods often focus on fitting abnormalities illustrated by the given anomaly examples only (i.e., seen anomalies), and consequently they fail to generalize to those that are not, i.e., new types/classes of anomaly unseen during training. To detect both seen and unseen anomalies, we introduce a novel deep weakly-supervised approach, namely Pairwise Relation prediction Network (PReNet), that learns pairwise relation features and anomaly scores by predicting the relation of any two randomly sampled training instances, in which the pairwise relation can be anomaly-anomaly, anomaly-unlabeled, or unlabeled-unlabeled. Since unlabeled instances are mostly normal, the relation prediction enforces a joint learning of anomaly-anomaly, anomaly-normal, and normal-normal pairwise discriminative patterns, respectively. PReNet can then detect any seen/unseen abnormalities that fit the learned pairwise abnormal patterns, or deviate from the normal patterns. Further, this pairwise approach also seamlessly and significantly augments the training anomaly data. Empirical results on 12 real-world datasets show that PReNet significantly outperforms nine competing methods in detecting seen and unseen anomalies. We also theoretically and empirically justify the robustness of our model w.r.t. anomaly contamination in the unlabeled data. The code is available at https://github.com/mala-lab/PReNet.

Understanding how topics within a document evolve over its structure is an interesting and important problem. In this paper, we address this problem by presenting a novel variant of Latent Dirichlet Allocation (LDA): Sequential LDA (SeqLDA). This variant directly considers the underlying sequential structure, i.e., a document consists of multiple segments (e.g., chapters, paragraphs), each of which is correlated to its previous and subsequent segments. In our model, a document and its segments are modelled as random mixtures of the same set of latent topics, each of which is a distribution over words; and the topic distribution of each segment depends on that of its previous segment, the one for first segment will depend on the document topic distribution. The progressive dependency is captured by using the nested two-parameter Poisson Dirichlet process (PDP). We develop an efficient collapsed Gibbs sampling algorithm to sample from the posterior of the PDP. Our experimental results on patent documents show that by taking into account the sequential structure within a document, our SeqLDA model has a higher fidelity over LDA in terms of perplexity (a standard measure of dictionary-based compressibility). The SeqLDA model also yields a nicer sequential topic structure than LDA, as we show in experiments on books such as Melville's "The Whale".

It is useful, sometimes crucial in medicine domain, to discover a temporal association or causal relationship among events. Such a mining problem is often challenging because ‘consequence events’ may not reliably occur after each trigger event of int

The work is motivated by real-world applications of detecting Adverse Drug Reactions (ADRs) from administrative health databases. ADRs are a leading cause of hospitalization and death worldwide. Almost all current postmarket ADR signaling techniques are based on spontaneous ADR case reports, which suffer from serious underreporting and latency. However, administrative health data are widely and routinely collected. They, especially linked together, would contain evidence of all ADRs. To signal unexpected and infrequent patterns characteristic of ADRs, we propose a domain-driven knowledge representation Unexpected Temporal Association Rule (UTAR), its interestingness measure, unexlev, and a mining algorithm MUTARA (Mining UTARs given the Antecedent). We then establish an improved algorithm, HUNT, for highlighting infrequent and unexpected patterns by comparing their ranks based on unexlev with those based on traditional leverage. Various experimental results on real-world data substantiate that both MUTARA and HUNT can signal suspected ADRs while traditional association mining techniques cannot. HUNT can reliably shortlist statistically significantly more ADRs than MUTARA (p=0.00078). HUNT, e.g., not only shortlists the drug alendronate associated with esophagitis as MUTARA does, but also shortlists alendronate with diarrhoea and vomiting for older ({\rm age} \ge 60) females. We also discuss signaling ADRs systematically by using HUNT.

It is well known that either domain specific or domain independent knowledge has been adopted in Information retrieval (IR) to improve the retrieval performance. In this paper, we propose a novel IR model for digital forensics by using latent semantic indexing (LSI) and WordNet as an underlying reference ontology to retrieve suspicious emails according to the semantic meaning of an investigatorpsilas query. Our model incorporates corpus independent knowledge from WordNet and corpus dependent knowledge from LSI into query expansion and reduction; and LSI is also adopted to simulate human meaning based judgement of relatedness between investigatorpsilas queries and emails. We compare the performance of the resulting LSI And WordNet based Information retrieval system (LAWIRS) with other three systems we implement, i.e. the LSI system, the Lucene system and the Lucene system with query expansion. Experimental results on several email datasets demonstrate that for short Boolean queries, LAWIRS can successfully capture their meaning and yield substantial improvements in the overall retrieval performance.

Spatial analysis in crime databases has recently been an active research topic. To solve the problem of finding the closest pairs of objects within a given spatial region, as required in crime geo-data applications, this paper proposes an efficient constrained k-closest pairs query processing algorithm based on growing window. It expands the window gradually instead of searching the whole workspace for multiple types of spatial objects. It employs a density-based range estimation approach to calculate the square query range and an optimized R-tree to store the index entities. In addition, a distance threshold T for the closest pair of objects is introduced to prune tree nodes. Experiments evaluate the effect of three important factors, i.e., the portion of overlapping between the workspaces of two data sets, the value of k, and the size of buffer. The results show that the new algorithm outperforms the heap-based approach.

In various real-world applications, it is very useful mining unanticipated episodes where certain event patterns unexpectedly lead to outcomes, e.g., taking two medicines together sometimes causing an adverse reaction. These unanticipated episodes are usually unexpected and infrequent, which makes existing data mining techniques, mainly designed to find frequent patterns, ineffective. In this paper, we propose unexpected temporal association rules (UTARs) to describe them. To handle the unexpectedness, we introduce a new interestingness measure, residual-leverage, and develop a novel case-based exclusion technique for its calculation. Combining it with an event-oriented data preparation technique to handle the infrequency, we develop a new algorithm MUTARC to find pairwise UTARs. The MUTARC is applied to generate adverse drug reaction (ADR) signals from real-world healthcare administrative databases. It reliably shortlists not only six known ADRs, but also another ADR, flucloxacillin possibly causing hepatitis, which our algorithm designers and experiment runners have not known before the experiments. The MUTARC performs much more effectively than existing techniques. This paper clearly illustrates the great potential along the new direction of ADR signal generation from healthcare administrative databases.

In this paper, we discuss a problem of finding risk patterns in medical data. We define risk patterns by a statistical metric, relative risk, which has been widely used in epidemiological research. We characterise the problem of mining risk patterns as an optimal rule discovery problem. We study an anti-monotone property for mining optimal risk pattern sets and present an algorithm to make use of the property in risk pattern discovery. The method has been applied to a real world data set to find patterns associated with an allergic event for ACE inhibitors. The algorithm has generated some useful results for medical researchers.

The scalability problem in data mining involves the development of methods for handling large databases with limited computational resources such as memory and computation time. In this paper, two scalable clustering algorithms, bEMADS and gEMADS, are presented based on the Gaussian mixture model. Both summarize data into subclusters and then generate Gaussian mixtures from their data summaries. Their core algorithm, EMADS, is defined on data summaries and approximates the aggregate behavior of each subcluster of data under the Gaussian mixture model. EMADS is provably convergent. Experimental results substantiate that both algorithms can run several orders of magnitude faster than expectation-maximization with little loss of accuracy.

The issue of obtaining a well-converged and well-distributed set of Pareto optimal solutions efficiently and automatically is crucial in multiobjective evolutionary algorithms (MOEAs). Many studies have proposed different evolutionary algorithms that can progress towards Pareto optimal sets with a wide-spread distribution of solutions. However, most mathematically convergent MOEAs desire certain prior knowledge about the objective space in order to efficiently maintain widespread solutions. We propose, based on our novel E-dominance concept, an adaptive rectangle archiving (ARA) strategy that overcomes this important problem. The MOEA with this archiving technique provably converges to well-distributed Pareto optimal solutions without prior knowledge. ARA complements the existing archiving techniques, and is useful to both researchers and practitioners.

The scalability problem in data mining involves the development of methods for handling large databases with limited computational resources. We present a two-phase scalable model-based clustering framework: first, a large data set is summed up into subclusters; Then, clusters are directly generated from the summary statistics of subclusters by a specifically designed expectation-maximization (EM) algorithm. Taking example for Gaussian mixture models, we establish a provably convergent EM algorithm, EMADS, which embodies cardinality, mean, and covariance information of each subcluster explicitly. Combining with different data summarization procedures, EMADS is used to construct two clustering systems: gEMADS and bEMADS. The experimental results demonstrate that they run several orders of magnitude faster than the classic EM algorithm with little loss of accuracy. They generate significantly better results than other model-based clustering systems using similar computational resources.

As a typical combinatorial optimization problem, the traveling salesman problem (TSP) has attracted extensive research interest. In this paper, we develop a self-organizing map (SOM) with a novel learning rule. It is called the integrated SOM (ISOM) since its learning rule integrates the three learning mechanisms in the SOM literature. Within a single learning step, the excited neuron is first dragged toward the input city, then pushed to the convex hull of the TSP, and finally drawn toward the middle point of its two neighboring neurons. A genetic algorithm is successfully specified to determine the elaborate coordination among the three learning mechanisms as well as the suitable parameter setting. The evolved ISOM (eISOM) is examined on three sets of TSP to demonstrate its power and efficiency. The computation complexity of the eISOM is quadratic, which is comparable to other SOM-like neural networks. Moreover, the eISOM can generate more accurate solutions than several typical approaches for TSP including the SOM developed by Budinich, the expanding SOM, the convex elastic net, and the FLEXMAP algorithm. Though its solution accuracy is not yet comparable to some sophisticated heuristics, the eISOM is one of the most accurate neural networks for the TSP.

The self-organizing map (SOM) is a powerful tool in the exploratory phase of data mining. However, due to the dimensional conflict, neighborhood preservation cannot always lead to perfect topology preservation. In this paper we establish an expanding SOM (ESOM) to detect and preserve better topology correspondence between the two spaces. Our experiment results demonstrate that the ESOM constructs better mappings than the classic SOM in terms of both topological and quantization errors. Furthermore, clustering results generated by the ESOM are more accurate than those of the SOM.