Ensemble fuzzy belief intrusion detection design

With the rapid growth of the Internet, computer attacks are increasing at a fast pace and can easily cause millions of dollar in damage to an organization. Detecting these attacks is an important issue of computer security. There are many types of attacks and they fall into four main categories, Denial of Service (DoS) attacks, Probe, User to Root (U2R) attacks, and Remote to Local (R2L) attacks. Within these categories, DoS and Probe attacks continuously show up with greater frequency in a short period of time when they attack systems. They are different from the normal traffic data and can be easily separated from normal activities. On the contrary, U2R and R2L attacks are embedded in the data portions of the packets and normally involve only a single connection. It becomes difficult to achieve satisfactory detection accuracy for detecting these two attacks. Therefore, we focus on studying the ambiguity problem between normal activities and U2R/R2L attacks. The goal is to build a detection system that can accurately and quickly detect these two attacks. In this dissertation, we design a two-phase intrusion detection approach. In the first phase, a correlation-based feature selection algorithm is proposed to advance the speed of detection. Features with poor prediction ability for the signatures of attacks and features inter-correlated with one or more other features are considered redundant. Such features are removed and only indispensable information about the original feature space remains. In the second phase, we develop an ensemble intrusion detection system to achieve accurate detection performance. The proposed method includes multiple feature selecting intrusion detectors and a data mining intrusion detector. The former ones consist of a set of detectors, and each of them uses a fuzzy clustering technique and belief theory to solve the ambiguity problem. The latter one applies data mining technique to automatically extract computer users’ normal behavior from training network traffic data. The final decision is a combination of the outputs of feature selecting and data mining detectors. The experimental results indicate that our ensemble approach not only significantly reduces the detection time but also effectively detect U2R and R2L attacks that contain degrees of ambiguous information.

Ensemble Fuzzy Belief Intrusion Detection Design

With the rapid growth of the Internet, computer attacks are increasing at a fast pace and can easily cause millions of dollar in damage to an organization. Detecting these attacks is an important issue of computer security. There are many types of attacks and they fall into four main categories, Denial of Service (DoS) attacks, Probe, User to Root (U2R) attacks, and Remote to Local (R2L) attacks. Within these categories, DoS and Probe attacks continuously show up with greater frequency in a short period of time when they attack systems. They are different from the normal traffic data and can be easily separated from normal activities. On the contrary, U2R and R2L attacks are embedded in the data portions of the packets and normally involve only a single connection. It becomes difficult to achieve satisfactory detection accuracy for detecting these two attacks. Therefore, we focus on studying the ambiguity problem between normal activities and U2R/R2L attacks. The goal is to build a detection system that can accurately and quickly detect these two attacks. In this dissertation, we design a two-phase intrusion detection approach. In the first phase, a correlation-based feature selection algorithm is proposed to advance the speed of detection. Features with poor prediction ability for the signatures of attacks and features inter-correlated with one or more other features are considered redundant. Such features are removed and only indispensable information about the original feature space remains. In the second phase, we develop an ensemble intrusion detection system to achieve accurate detection performance. The proposed method includes multiple feature selecting intrusion detectors and a data mining intrusion detector. The former ones consist of a set of detectors, and each of them uses a fuzzy clustering technique and belief theory to solve the ambiguity problem. The latter one applies data mining technique to automatically extract computer users’ normal behavior from training network traffic data. The final decision is a combination of the outputs of feature selecting and data mining detectors. The experimental results indicate that our ensemble approach not only significantly reduces the detection time but also effectively detect U2R and R2L attacks that contain degrees of ambiguous information.

Effects of shading on calcareous benthic periphyton in a short-hydroperiod oligotrophic wetland (Everglades, FL, USA)

The effects of shade on benthic calcareous periphyton were tested in a short-hydroperiod oligotrophic subtropical wetland (freshwater Everglades). The experiment was a split-plot design set in three sites with similar environmental characteristics. At each site, eight randomly selected 1-m2 areas were isolated individually in a shade house, which did not spectrally change the incident irradiance but reduced it quantitatively by 0, 30, 50, 60, 70, 80, 90 and 98%. Periphyton mat was sampled monthly under each shade house for a 5 month period while the wetland was flooded. Periphyton was analyzed for thickness, DW, AFDW, chlorophyll a (chl a) and incubated in light and dark BOD bottles at five different irradiances to assess its photosynthesis–irradiance (PI) curve and respiration. The PI curves parameters P max, I k and eventually the photoinhibition slope (β) were determined following non-linear regression analyses. Taxonomic composition and total algal biovolume were determined at the end of the experiment. The periphyton composition did not change with shade but the PI curves were significantly affected by it. I k increased linearly with increasing percent irradiance transmittance (%IT = 1−%shade). P max could be fitted with a PI curve equation as it increased with %IT and leveled off after 10%IT. For each shade level, the PI curve was used to integrate daily photosynthesis for a day of average irradiance. The daily photosynthesis followed a PI curve equation with the same characteristics as P max vs. %IT. Thus, periphyton exhibited a high irradiance plasticity under 0–80% shade but could not keep up the same photosynthetic level at higher shade, causing a decrease in daily GPP at 98% shade levels. The plasticity was linked to an increase in the chl a content per cell in the 60–80% shade, while this increase was not observed at lower shade likely because it was too demanding energetically. Thus, chl a is not a good metric for periphyton biomass assessment across variously shaded habitats. It is also hypothesized that irradiance plasticity is linked to photosynthetic coupling between differently comprised algal layers arranged vertically within periphyton mats that have different PI curves.

Effects of Simulated Drought on the Carbon Balance of Everglades Short-Hydroperiod Marsh

Hydrology drives the carbon balance of wetlands by controlling the uptake and release of CO2 and CH4. Longer dry periods in between heavier precipitation events predicted for the Everglades region, may alter the stability of large carbon pools in this wetland's ecosystems. To determine the effects of drought on CO2 fluxes and CH4 emissions, we simulated changes in hydroperiod with three scenarios that differed in the onset rate of drought (gradual, intermediate, and rapid transition into drought) on 18 freshwater wetland monoliths collected from an Everglades short-hydroperiod marsh. Simulated drought, regardless of the onset rate, resulted in higher net CO2 losses net ecosystem exchange (NEE) over the 22-week manipulation. Drought caused extensive vegetation dieback, increased ecosystem respiration (Reco), and reduced carbon uptake gross ecosystem exchange (GEE). Photosynthetic potential measured by reflective indices (photochemical reflectance index, water index, normalized phaeophytinization index, and the normalized difference vegetation index) indicated that water stress limited GEE and inhibited Reco. As a result of drought-induced dieback, NEE did not offset methane production during periods of inundation. The average ratio of net CH4 to NEE over the study period was 0.06, surpassing the 100-year greenhouse warming compensation point for CH4 (0.04). Drought-induced diebacks of sawgrass (C3) led to the establishment of the invasive species torpedograss (C4) when water was resupplied. These changes in the structure and function indicate that freshwater marsh ecosystems can become a net source of CO2 and CH4 to the atmosphere, even following an extended drought. Future changes in precipitation patterns and drought occurrence/duration can change the carbon storage capacity of freshwater marshes from sinks to sources of carbon to the atmosphere. Therefore, climate change will impact the carbon storage capacity of freshwater marshes by influencing water availability and the potential for positive feedbacks on radiative forcing.