Lon Mu Liu

Data mining on time series: an illustration using fast-food restaurant franchise data

Abstract Given the widespread use of modern information technology, a large number of time series may be collected during normal business operations. We use a fast-food restaurant franchise as a case to illustrate how data mining can be applied to such time series, and help the franchise reap the benefits of such an effort. Time series data mining at both the store level and corporate level are discussed. Box–Jenkins seasonal ARIMA models are employed to analyze and forecast the time series. Instead of a traditional manual approach of Box–Jenkins modeling, an automatic time series modeling procedure is employed to analyze a large number of highly periodic time series. In addition, an automatic outlier detection and adjustment procedure is used for both model estimation and forecasting. The improvement in forecast performance due to outlier adjustment is demonstrated. Adjustment of forecasts based on stored historical estimates of like-events is also discussed. Outlier detection also leads to information that can be used not only for better inventory management and planning, but also to identify potential sales opportunities. To illustrate the feasibility and simplicity of the above automatic procedures for time series data mining, the SCA Statistical System is employed to perform the related analysis.

Impact of methamphetamine precursor chemical legislation, a suppression policy, on the demand for drug treatment

Research is needed to help treatment programs plan for the impacts of drug suppression efforts. Studies to date indicate that heroin suppression may increase treatment demand. This study examines whether treatment demand was impacted by a major US methamphetamine suppression policy -- legislation regulating precursor chemicals. The precursors ephedrine and pseudoephedrine, in forms used by large-scale methamphetamine producers, were regulated in August 1995 and October 1997, respectively. ARIMA-intervention time-series analysis was used to examine the impact of each precursors regulation on monthly voluntary methamphetamine treatment admissions (a measure of treatment demand), including first-time admissions and re-admissions, in California (1992-2004). Cocaine, heroin, and alcohol treatment admissions were used as quasi-control series. The 1995 regulation of ephedrine was found to be associated with a significant reduction in methamphetamine treatment admissions that lasted approximately 2 years. The 1997 regulation of pseudoephedrine was associated with a significant reduction that lasted approximately 4 years. First-time admissions declined more than re-admissions. Cocaine, heroin, and alcohol admissions were generally unaffected. While heroin suppression may be associated with increased treatment demand as suggested by research to date, this study indicates that methamphetamine precursor regulation was associated with decreases in treatment demand. A possible explanation is that, during times of suppression, heroin users may seek treatment to obtain substitute drugs (e.g., methadone), while methamphetamine users have no comparable incentive. Methamphetamine suppression may particularly impact treatment demand among newer users, as indicated by larger declines in first-time admissions.

Methamphetamine suppression and route of administration: precursor regulation impacts on snorting, smoking, swallowing and injecting

AIMS The route of drug administration affects risk for dependence and medical harm. This study examines whether routes used by methamphetamine treatment participants were impacted by a major drug suppression policy-federal regulation of the methamphetamine precursor chemicals ephedrine and pseudoephedrine. DESIGN Autoregressive-integrated moving average (ARIMA) intervention time-series analysis. SETTING California (1992-2004). INTERVENTIONS Ephedrine single-ingredient products regulation, implemented August 1995; ephedrine with other active medicinal ingredients regulation, implemented October 1996; pseudoephedrine products regulation, implemented October 1997. MEASUREMENTS Monthly counts of non-coerced methamphetamine treatment admissions reporting snorting, smoking, swallowing or injecting. Findings After rising sharply, snorting, smoking, swallowing and injecting admissions dropped 50%, 43%, 26% and 26%, respectively, when the 1995 regulation was implemented. Snorting also dropped 38% at the time of the 1997 regulation. Snorting, swallowing and injecting remained at lower levels to the end of the study period. Smoking resurged (40%) at the time of the 1996 regulation and continued rising. CONCLUSIONS Precursor regulation was associated with changes in the administration of methamphetamine. Injecting, the route with the greatest health risk, entered a long-term reduction. So, too, did snorting and swallowing, two routes with lower risk for dependence. In contrast, smoking, which has a relatively high risk for dependence, dropped, then rebounded and entered a long-term rise. A possible explanation is that injecting, snorting and swallowing were largely linked with US domestic methamphetamine production, which has yet to recover from the regulations. While Mexican production, which was impacted only temporarily by the regulations and has supplanted domestic production, may have helped to diffuse smoking, a route with which it is historically correlated.

Mexico's precursor chemical controls: Emergence of less potent types of methamphetamine in the United States

BACKGROUND This study examines whether Mexicos controls on ephedrine and pseudoephedrine, the two precursor chemicals that yield the most potent form of methamphetamine, d-methamphetamine, impacted the prevalence/availability of less potent types of methamphetamine in the United States-types associated with the alternative precursor chemical P2P. METHOD Using ARIMA-intervention time series analysis of monthly drug exhibits (a prevalence/availability indicator) from the System to Retrieve Information from Drug Evidence (STRIDE), we tested whether Mexicos controls, which began in 2005, were associated with growth/decline in d-methamphetamine and growth/decline in P2P-associated, less potent l-methamphetamine, racemic methamphetamine (a 50:50 ratio of d- and l-isomers), and mixed isomer methamphetamine (an unequal ratio of d- and l-isomers). Heroin, cocaine and marijuana exhibits were used for quasi-control (01/2000-04/2011). RESULTS Mixed-isomer exhibits constituted about 4% of the methamphetamine exhibits before Mexicos controls, then rose sharply in association with them and remained elevated, constituting about 37% of methamphetamine exhibits in 2010. d-Methamphetamine exhibits dropped sharply; l-methamphetamine and racemic methamphetamine exhibits had small rises. d-Methamphetamine exhibits partially recovered in the US West, but little recovery occurred in the US Central/South. Quasi-control series were generally unaffected. CONCLUSION The US methamphetamine market changed. Widespread emergence of less potent methamphetamine occurred in conjunction with Mexicos controls. And prevalence/availability of the most potent type of the drug, d-methamphetamine, declined, a partial recovery in the West notwithstanding. Granting that lower potency drugs typically engender less dependence and attendant problems, these findings suggest that, following Mexicos controls, the potential harm of a sizeable amount of the US methamphetamine supply decreased.

US federal cocaine essential (‘precursor’) chemical regulation impacts on US cocaine availability: an intervention time–series analysis with temporal replication

Abstract Background and Aims Research shows that essential/precursor chemical controls have had substantial impacts on US methamphetamine and heroin availability. This study examines whether US federal essential chemical regulations have impacted US cocaine seizure amount, price and purity—indicators of cocaine availability. Design Autoregressive integrated moving average (ARIMA)‐intervention time–series analysis was used to assess the impacts of four US regulations targeting cocaine manufacturing chemicals: potassium permanganate/selected solvents, implemented October 1989 sulfuric acid/hydrochloric acid, implemented October 1992; methyl isobutyl ketone, implemented May 1995; and sodium permanganate, implemented December 2006. Of these chemicals, potassium permanganate and sodium permanganate are the most critical to cocaine production. Setting Conterminous United States (January 1987—April 2011). Measurements Monthly time–series: purity‐adjusted cocaine seizure amount (in gross weight seizures < 6000 grams), purity‐adjusted price (all available seizures), and purity (all available seizures). Data source: System to Retrieve Information from Drug Evidence. Findings The 1989 potassium permanganate/solvents regulation was associated with a seizure amount decrease (change in series level) of 28% (P < 0.05), a 36% increase in price (P < 0.05) and a 4% decrease in purity (P < 0.05). Availability recovered in 1–2 years. The 2006 potassium permanganate regulation was associated with a 22% seizure amount decrease (P < 0.05), 100% price increase (P < 0.05) and 35% purity decrease (P < 0.05). Following the 2006 regulation, essentially no recovery occurred to April 2011. The other two chemical regulations were associated with statistically significant but lesser declines in indicated availability. Conclusions In the United States, essential chemical controls from 1989 to 2006 were associated with pronounced downturns in cocaine availability.

Hazardous birthday drinking among young people: population‐based impacts on emergency department and in‐patient hospital admissions

BACKGROUND AND AIMS There is growing concern about the possible adverse health impacts of binge drinking during birthday celebrations among adolescents and young adults. We estimate the impacts of birthday alcohol use on adolescent and young adult in-patient/emergency department (ED) hospital admissions. DESIGN We employed Autoregressive Integrated Moving Average (ARIMA) intervention analysis to assess whether the rate of ICD-10 alcohol-use-disorder (AUD) events per 1000 in-patient/ED admissions increased significantly during birthday weeks. SETTING All in-patient/ED admissions in Ontario, Canada from 1 April 2002 to 31 March 2007. PARTICIPANTS Individuals aged 12-30 years. MEASUREMENTS AUD events per 1000 in-patient/ED admissions by age in weeks. FINDINGS Multiple increases were found. The largest occurred during the birthday week of 19 years of age, the beginning of the minimum legal drinking age (MLDA) in Ontario: AUD admission rates increased (spiked) by 38.30 per 1000 total admissions [95% confidence interval (CI) = 34.66-41.94] among males (a 114.3% increase over baseline), and by 28.13 (95% CI = 25.56-30.70) among females (a 164.0% increase). Among both genders, the second largest birthday-week spikes occurred during ages 20-22 years, followed by somewhat lower but still pronounced birthday-week spikes during ages 23-26 years and 30 years (all these spikes: P < 0.05). Birthday-week spikes occurred as early as age 16 years for males and 14 years for females (both spikes: P < 0.05). CONCLUSIONS There appears to be an increase in alcohol-related adverse events from drinking around the time of ones birthday among young adults in Canada.