Churn prediction is used to identify customers that are becoming less loyal and is an important tool for companies that want to stay competitive in a rapidly growing market. In retail, a dynamic definition of churn is needed to identify churners correctly. Customer Lifetime Value (CLV) is the monetary value of a customer relationship. No change in CLV for a given customer indicates a decrease in loyalty. This thesis proposes a novel approach to churn prediction. The proposed model uses a Recurrent Neural Network to identify churners based on Customer Lifetime Value time series regression. The results show that the model performs better than random. This thesis also investigated the use of the K-means algorithm as a replacement to a rule-extraction algorithm. The K-means algorithm contributed to a more comprehensive analytical context regarding the churn prediction of the proposed model. / Illojalitet prediktering används för att identifiera kunder som är påväg att bli mindre lojala och är ett hjälpsamt verktyg för att ett företag ska kunna driva en konkurrenskraftig verksamhet. I detaljhandel behöves en dynamisk definition av illojalitet för att korrekt kunna identifera illojala kunder. Kundens livstidsvärde är ett mått på monetärt värde av en kundrelation. En avstannad förändring av detta värde indikerar en minskning av kundens lojalitet. Denna rapport föreslår en ny metod för att utföra illojalitet prediktering. Den föreslagna metoden består av ett återkommande neuralt nätverk som används för att identifiera illojalitet hos kunder genom att prediktera kunders livstidsvärde. Resultaten visar att den föreslagna modellen presterar bättre jämfört med slumpmässig metod. Rapporten undersöker också användningen av en k-medelvärdesalgoritm som ett substitut för en regelextraktionsalgoritm. K-medelsalgoritm bidrog till en mer omfattande analys av illojalitet predikteringen.

Wireless Sensor Networks (WSNs) experiences several technical challenges such as limited energy, short transmission range, limited storage capacities, and limited computational capabilities. Moreover, the sensor nodes are deployed randomly and massively over an inaccessible or hostile region. Hence, WSNs are vulnerable to adversaries and are usually operated in a dynamic and unreliable environment. Animal tracking using wireless sensors is one such application of WSN where power management plays a vital role. In this paper, an energy-efficient hybrid routing method is proposed that divides the whole network into smaller regions based on sensor location and chooses the routing scheme accordingly. The sensor network consists of a base station (BS) located at a distant place outside the network, and a relay node is placed inside the network for direct communications from nodes nearer to it. The nodes are further divided into two categories based on the supplied energy; such that the ones located far away from BS and relay have higher energy than the nodes nearer to them. The network performance of the proposed method is compared with protocols like LEACH, SEP, and SNRP, considering parameters like stability period, throughput and energy consumption. Simulation result shows that the proposed method outperforms other methods with better network performance.

Thermal barrier coatings (TBCs) have been widely used in gas turbine engines to protect the underlying metal from high operating temperature so as to improve the durability of the components and enhance the engine efficiency. However, since the TBCs always operate in a demanding high-temperature environment of aircraft and industrial gas-turbine engines, a better understanding of this complex system is required to improve the durability and reliability. The objective of this study is to investigate the effects of surface modification for the NiCoCrAlY bond coats on the thermal cycling lifetime and failure characteristics of TBCs. Parameters of modification for the bond coats included as-sprayed, barrel-finished, hand-polished and pre-oxidation heat treatment at 1100[degrees]C in P=10O2-8 atm up to 4 hours, carried out prior to the electron beam physical vapor deposition (EB-PVD) of ZrO2-7wt% Y2O3 (7YSZ) ceramic topcoat. The resulting characteristics of the bond coat and the thermally grown oxide (TGO) scale were initially documented by surface roughness, phase constituents of the TGO scale, and residual stress of the TGO scale. The thermal cycling test consisted of 10-minute heat-up to 1121°C, 40-minute hold at 1121°C, and 10-minute forced air-quench. As-coated and thermally-cycled TBCs were characterized by optical profilometry (OPM), photo-stimulated luminescence spectroscopy (PSLS), optical microscopy, scanning electron microscopy (SEM) equipped with energy dispersive spectroscopy (EDS), and scanning/transmission electron microscopy (TEM/STEM) equipped with high angle annular dark field (HAADF) and X-ray energy dispersive spectroscopy (XEDS). TBC specimens for TEM/STEM analysis were prepared by focused ion beam (FIB) in-situ lift-out (INLO) technique. Superior thermal cycling lifetime was observed for TBCs with as-sprayed bond coats regardless of pre-oxidation heat treatment, and TBCs with hand-polished bond coats only after pre-oxidation heat treatment. With pre-oxidation heat treatment, relative photostimulated luminescence intensity of the equilibrium α-Al2O3 increased. Thus, the improvement in TBC lifetime can be correlated with an increase in the amount of α-Al2O3 in the TGO scale, given a specific surface modification/roughness. The lifetime improvement due to pre-oxidation was particularly significant to TBCs with smooth hand-polished bond coats and negligible for TBCs with rough as-sprayed bond coats. Spallation-fracture paths depended on the lifetime of TBCs. Premature spallation of TBCs occurred at the interface between the YSZ and TGO. Longer durability can be achieved by restricting the fracture paths to the TGO/bond coat interface. Small particulate phase observed through the TGO scale was identified as Y2O3 (cubic) by diffraction analysis on TEM. While small addition of Y in the NiCoCrAlY bond coat helps the adhesion of the TGO scale, excessive alloying can lead to deleterious effects.

Minority carrier transport properties and the effects of electron irradiation/injection were studied in GaN and ZnO containing dopants known to form acceptor states deep within the materials' bandgap. Minority carrier diffusion length and lifetime changes were investigated using Electron Beam Induced Current (EBIC) method, cathodoluminescence spectroscopy, spectral photoresponse and persistent photoconductivity measurements. It is shown that electron irradiation by the beam of a scanning electron microscope results in a significant increase of minority carrier diffusion length. These findings are supported by the cathodoluminescence measurements that demonstrate the decay of near-band-edge intensity as a consequence of increasing carrier lifetime under continuous irradiation by the electron beam. Temperature-dependent measurements were used to determine the activation energies for the electron irradiation-induced effects. The latter energies were found to be consistent with the involvement of deep acceptor states. Based on these findings, the effects of electron irradiation are explained via the mechanism involving carrier trapping on these levels. Solid-state electron injection was also shown to result in a similar increase of minority carrier lifetime and diffusion length. Solid-state injection was carried out by applying the forward bias to a ZnO homojunction and resulted in a significant improvement of the peak photoresponse of the junction. This improvement was unambiguously correlated with the increase of the minority carrier diffusion length due to electron injection.

This study investigates the factors that affect the lifetime and popping location of surface bubbles. The experiment was conducted using three different liquids (water, Sodium Dodecyl Sulfate, and Decane) with varying bubble sizes, using three different needle sizes. Each setup was tested 50 times. For pure water bubbles, the foot of the bubble is the most critical location because it typically has the highest temperature gradient, which creates a localized Marangoni flow that thins the film and eventually leads to the bubble bursting at the foot. When SDS was added to water, the bubble lifetime increased significantly. This is because the Marangoni stresses were reduced, and the bubble film thinned mainly due to curvature-driven drainage flow. The lifetime of the SDS bubble had a positive correlation with increasing bubble size. For Decane bubbles, the volatility of the liquid plays a significant role in the lifetime and popping location of the bubble. When the Decane was heated to 40°C, the lifetime of the bubbles increased significantly from 0-20 seconds to 8-12 minutes. This is because the high volatility of the Decane caused rapid evaporation of the bubble cap at the interface, which cooled the surface of the liquid. This temperature difference creates a difference in surface tension, which causes the liquid to flow from the bulk liquid into the apex of the bubble, thickening the cap film until it cools down. Then, it pops from the top due to the curvature-driven drainage.

These studies deal with lifetime profit analyses for individual cows, and using these estimates to determine the economic value of genetic changes in traits for which genetic evaluations (predicted transmitting ability, PTA) are currently available. Data were collected from six states processed by Dairy Records Management Systems (DRMS) from herds on continuous test for at least 10 yr. The purpose of the first study was to determine how well estimators of lifetime net income based on 305-d lactation yields and a 10-yr opportunity (RNI305.10) and based on complete lactation data but a 5-yr opportunity (RNIc.5) predict the estimate based on complete lactations and a 10-yr opportunity (RNIc.10). Records for 22,854 cows in Virginia herds born in 1988, 1990, and 1992 from the DRMS in Raleigh, NC were used. Each RNI was calculated using fluid (skim/fat) pricing and milk-fat-protein pricing. Regression analyses including herd and birth year were used in the model to estimate the regression of RNIc.10 on RNIc.5, and RNIc.10 on RNI305.10. The resulting regression coefficients for fluid (skim/fat) pricing were $1.53 and $1.12 explaining 67 and 97% of the variation of RNIc.10, respectively. The corresponding results for milk-fat-protein pricing were $1.52 and $1.14 explaining 68 and 96% of the variation of RNIc.10, respectively. Using RNIc.10 as the measure to estimate lifetime profit is strongly recommended over the two alternatives tested. In the second study, the economic impacts of a unit change in PTA of daughter pregnancy rate (DPR) and other PTA in the merit indexes on lifetime profit estimates of a bull's daughters were estimated to determine an economic weight for the PTADPR and other PTA in economic indexes. Records for 71,094 cows born in 1988, 1990, and 1992 from six states processed at DRMS were used: Florida [10,940 cows], Indiana [8,231 cows], North Carolina [12,280 cows], Texas [4,786 cows], Virginia [20,341 cows], and Vermont [14,516 cows]. The basic RNI function consisted of [total milk, fat, and protein income ?feed cost for production] (yield income, YI) + [net value of calves + net salvage value] (non yield income, NYI) ?rearing cost (RC) ?[(daily cost for labor, maintenance feed, supplies, and fixed expenses) x days in herd] (daily cost, DC). Some of the economic impacts of PTA described for the merit indexes were not included in the basic RNI. These were added to RNI by multiplying the respective sire PTA by the economic impact. These included -165*PTASCS (M); 33*udder composite + 15*feet and legs composite -14.86*body size composite (T); and 8.064*PTA for daughter pregnancy rate -4.80*PTA for daughter calving ease (PRCE). Each ARNI was calculated using all production records initiated prior to the cow's tenth birthday with three milk pricing systems comparable to the prices in USDA three merit indexes: fluid (skim/fat) pricing (FARNI), milk-fat-protein pricing (NMARNI), and cheese pricing (CARNI). Two levels of prices for rearing cost per day and daily cost were used for calculating FARNI, NMARNI, and CARNI. Regression analyses including herd and birth year in the model were used to estimate the simple and partial regressions of ARNI or partitioned ARNI on sire PTA. Partial regression included all PTA in Net Merit, except service sire calving ease. Ignoring other PTA, one unit increase in PTADPR increased 476.25kg of lifetime total milk or 18 days of total DIM. One unit decrease in PTASCS increased 4372.50kg of lifetime total milk. With low daily and rearing costs, each 1% change in PTADPR increased ARNI by $59.31 to $55.82 depending on the milk pricing systems. The corresponding results with high daily and rearing costs were $27.50 to $24.01. Standardized multiple regression enabled the comparison of the economic weights of this study with those of USDA. The PTA for productive life (PL) in all three USDA merit index was emphasized less than the results from this study; however, PTADPR in USDA indexes was emphasized more than this study. In this study, the economic weight of PTADPR was negative within the low daily and rearing costs, but it was positive in the high daily and rearing costs. / Master of Science

Mellan 1961–1974 byggdes det knappt en halv miljon småhus i Sverige. Småhusen står idag inför omfattande renoveringar på grund av ålder och rådande byggteknik.Småhusägarna saknar ofta en teknisk underhållsplan vilket förkortar byggnadsdelars livslängder och ger ökade förvaltningskostnader.Rapporten består av en litteraturstudie som sammanställer problem och tillhörande åtgärder för småhus från denna period. Intervjuer och enkätundersökningar ligger till grund för framtagandet av handlingsplan för renovering och en teknisk underhållsplan.Med ökad kunskap kan lämpliga renoveringar utföras och samordnas vilket leder till lägre renoveringskostnader. En underhållsplan åskådliggör kostnader och visar när och hur underhåll bör utföras. Detta resulterar i lägre förvaltningskostnader samt förlängda tekniska livslängder på byggnadsdelar.Resultatet består av en sammanfattning av problem och lämpliga åtgärder. En handlingsplan som visar vilka renoveringar som bör utföras och hur de kan samordnas. En allmängiltig och lättöverskådlig teknisk underhållsplan. Sammantaget kan detta leda till ett mer hållbart bostadsbestånd ur ett ekonomiskt och ekologiskt perspektiv. / Between 1961-1974, almost half a million detached houses were built in Sweden. These detached houses will be undergoing extensive renovations due to their age and prevailing construction technology. Small house owners often lack a technical maintenance plan, which shortens the life of building components and increases management costs.The report consists of a literature study that compiles problems and related measures for detached houses from this period. Interviews and surveys form the basis for the development of an action plan for renovation and a technical maintenance plan.With increased knowledge, suitable renovations can be carried out and coordinated, which leads to lower renovation costs. A maintenance plan illustrates costs and shows when and how maintenance should be performed. This results in lower management costs and extends technical life of building components.The result consists of a summary of problems and appropriate measures. An action plan that shows which renovations should be carried out and how they can be coordinated. A universal and easy-to-understand technical maintenance plan. All in all, this can lead to a more sustainable housing stock from an economic and ecological perspective.

In recent years, the direct-to-consumer business (DTC) has become an increasingly popular way for companies to conduct their business. A large part of why companies choose to build a business based on DTC is for the consumer data they get access to. This study aims to understand how small emerging DTC companies achieve sustainable growth as they both have to acquire new customers and retain existing customers. The empirical data consisted of qualitative interviews with three DTC companies operating in three different industries. The result of this study, which was based on the dynamic marketing capabilities model, is that a DTC company best achieves sustainable growth from a customer-centric approach, where decisions are based on data. This allows the organization to work more agile and, as a result, proactive manner. / Under de senaste åren har man kunnat se ett skifte i hur allt fler direct-to-consumer (DTC)-företag har startat. Anledningen till att DTC har blivit ett populärt sätt att bedriva sin verksamhet genom är för att företagen kan nå ut till sina kunder på ett effektivt sätt och på så sätt kunna anpassa sin verksamhet efter dem. Syftet med denna studie är att förstå hur små framväxande DTC-företag uppnår hållbar tillväxt då de både måste anskaffa nya kunder och värna om redan existerande kunder. Den empiriska data bestod av kvalitativa intervjuer med tre DTC-företag, verksamma inom tre olika industrier. Resultatet av denna studie som utgick från dynamic marketing capabilities-modellen, är att ett DTC-företag bäst uppnår en hållbar tillväxt från ett kundcentrerat tillvägagångssätt, där beslut baseras på data. Genom tydliga och informativa processer möjliggör det för företag att jobba mer proaktivt.

No description available.

No description available.