Minimizing False Positives of a Decision Tree Classifier for Intrusion Detection on the Internet

Machine learning or data mining technologies are often used in network intrusion detection systems. An intrusion detection system based on machine learning utilizes a classifier to infer the current state from the observed traffic attributes. The problem with learning-based intrusion detection is that it leads to false positives and so incurs unnecessary additional operation costs. This paper investigates a method to decrease the false positives generated by an intrusion detection system that employs a decision tree as its classifier. The paper first points out that the information-gain criterion used in previous studies to select the attributes in the tree-constructing algorithm is not effective in achieving low false positive rates. Instead of the information-gain criterion, this paper proposes a new function that evaluates the goodness of an attribute by considering the significance of error types. The proposed function can successfully choose an attribute that suppresses false positives from the given attribute set and the effectiveness of using it is confirmed experimentally. This paper also examines the more trivial leaf rewriting approach to benchmark the proposed method. The comparison shows that the proposed attribute evaluation function yields better solutions than the leaf rewriting approach.

Copolymerization of Vinyl Chloride and Ethylene with Cp<Superscript>*</Superscript>Ti( OCH<Subscript>3</Subscript>)<Subscript>3</Subscript>/MAO Catalyst

Summary Copolymerization of vinyl chloride (VC) and ethylene with Cp^*Ti( OCH₃)₃/MAO catalyst was investigated. The Cp^*Ti( OCH₃)₃/MAO catalyst initiated the copolymerization of VC with ethylene, although the copolymer yields were low. In the ¹³C NMR spectra of the copolymers, the peaks based on junction part between VC and ethylene was observed, but the signals were small. From DSC measurement of the copolymers, only one glass transition temperature was observed. Thus, it is clear that the copolymerization with Cp^*Ti( OCH₃)₃/MAO catalysts gave the copolymer, and the copolymer consisting of block sequence rather than random copolymer. Received: 13 November 2002/Revised: 6 January 2003/Accepted: 10 January 2003 Correspondence to Kiyoshi Endo     

Grain refinement and superplasticity in a magnesium alloy processed by equal-channel angular pressing

The extrusion/equal channel angular pressing (EX-ECAP) processing procedure, in which magnesium-based alloys are subjected to extrusion followed by ECAP, was applied to a Mg-7.5 pct Al-0.2 pct Zr alloy prepared by casting. Microstructural inspection showed the EX-ECAP process was effective in reducing the grain size from ∼21 μm after extrusion to an as-pressed grain size of ∼0.8 μm. It is shown through static annealing that these ultrafine grains are reasonably stable up to 473 K, but grain growth occurs at higher temperatures. Tensile specimens were cut from the billets prepared by EX-ECAP and testing showed these specimens exhibited superplasticity at relatively low temperatures with maximum elongations up to >700 pct. By processing through EX-ECAP to a higher imposed strain and thereby increasing the area fraction of high-angle boundaries, it is demonstrated that there is a potential for achieving high-strain-rate superplasticity. This article is based on a presentation made at the Symposium entitled “Phase Transformations and Deformation in Magnesium Alloys,” which occurred during the Spring TMS meeting, March 14–18, 2004, in Charlotte, NC, under the auspices of the ASM-MSCTS Phase Transformations Committee.     

Achieving enhanced ductility in a dilute magnesium alloy through severe plastic deformation

Experiments were conducted to evaluate the utility of a new processing procedure developed for Mg-based alloys in which samples are subjected to a two-step processing route of extrusion followed by equal-channel angular pressing (designated as EX-ECAP). The experiments were conducted using a Mg-0.6 wt pct Zr alloy and, for comparison purposes, samples of pure Mg. It is shown that the potential for successfully using ECAP increases in both materials when adopting the EX-ECAP procedure. For the Mg-Zr alloy, the use of EX-ECAP produces a grain size of ∼1.4 μm when the pressing is undertaken at 573 K. By contrast, using EX-ECAP with pure Mg at 573 K produces a grain size of ∼26 μm. Tensile testing of the Mg-Zr alloy at 523 and 573 K after processing by EX-ECAP revealed the occurrence of significantly enhanced ductilities with maximum elongations of ∼300 to 400 pct.     

Connection‐Strength Estimation of Neuronal Networks by Fitting for Izhikevich Model

Recently, there has been abundant research using multineuron recording, but there are many problems with extracting the features from the obtained spike time series, which are huge in volume and complex. Here we introduce a new method of estimating synaptic connection strengths between neurons by fitting to the Izhikevich model by maximum likelihood estimation. We demonstrate that our method can estimate connection strengths from spike time series given by a simulated neural ensemble and can estimate nonconnectivity between two independent cultured neuronal networks. These results suggest that our method is applicable to network and plasticity analysis of neuronal networks.     

Research and Development of the Coprecipitation Process for Lanthanum Germanate Oxyapatite

Although lanthanum germanate oxyapatite (La–Ge–O) has shown good potential for use as a solid electrolyte in energy storage applications, its synthesis has been challenging by either solid‐ or solution‐state methods. In this study, a new synthesis of La–Ge–O was developed through a coprecipitation technique, in which a highly concentrated homogeneous aqueous solution of La and Ge was prepared from aqueous ammonium germanate and lanthanum nitrate solutions with the addition of dilute nitric acid. Several precipitates were formed by pH manipulation, including an amorphous material obtained at pH > 3. Compared to the individual precipitation behaviors of the parent compounds, the amorphous precipitate was formed only from the aqueous two‐component mixture, and appeared to contain both metals. This material was transformed into crystalline mixtures upon heating at 1273 K. The crystalline phases were La₂Ge₃O₉ and hexagonal‐type GeO₂ when the precipitate was formed below pH 8, and the La–Ge–O and La₂Ge₂O₇ phases when the precipitate was formed around pH 8. Product formation from the coprecipitate was discussed based on X‐ray diffraction and thermal analyses. The improved availability of La–Ge–O will allow more extensive investigations of its useful properties.     

Relationship between electron sensitivity and chemical structures of polymers as electron beam resists. VI: Electron sensitivity of various acetalized poly(vinyl alcohol)s

Acetalized poly(vinyl alcohol)s which were synthesized from poly(vinyl alcohol) (PVA) and aldehyde or ketone were evaluated as electron beam (EB) resists, in order to investigate the relationship between EB sensitivity and chemical structures of the polymers. It was found that the acetalized PVAs were easily crosslinked by EB exposure. The main mechanism of crosslinking may be radical reaction at acetal group. The sensitivity of the acetalized PVA depended on the structure of the acetal group. Acetalized PVAs synthesized from benzaldehyde, in which the electron attracting group was substituted on the benzene ring and from linear aliphatic aldehyde having a long alkyl chain, had high sensitivity. A high sensitivity of 7.2 × 10⁻⁷ C/cm² was attained when the acetalized PVA synthesized from p-chloro-benzaldehyde was used. Acetalized PVAs synthesized from PVA and aldehyde having a cyclic structure had an excellent dry etching durability and were suitable as negative EB resists.     

Effects of processing and cooking on the levels of pesticide residues in soybean samples

The effects of processing and cooking on the levels of pesticide residues in soybean samples were investigated for 14 pesticides in pre-harvest samples. On soaking, the transfer ratios (%, total pesticide residue amount in product/that in soybean) of soaked soybean were greater than 60% for most of the pesticides investigated. The transfer ratio of soymilk ranged from 37% to 92%, and that of tofu ranged from 7% to 63%. The processing factor (Pf, the concentration (mg/kg) of pesticide in product/that in soybean) of tofu ranged from 0.026 to 0.28. These values varied among pesticides. There was a high correlation between the log P(ow) and the transfer ratio of tofu. The test described here should be useful to obtain the transfer ratios of pesticide residues in processing and/or cooking steps.     

Effects of fatty acid and emulsifier on the complex formation and in vitro digestibility of gelatinized potato starch

The effects of fatty acid, monoacylglycerol, and polyglycerol fatty acid ester with varying chain length in their acyl chains on the extent of complex formation (complex index) and in vitro enzymatic digestibility of gelatinized potato starch were investigated. The complex index increased with increase in the concentration of the ligands (fatty acid, monoacylglycerol, and polyglycerol fatty acid ester), with the plateau in the complex index value depending on the type of ligands. In comparison of complex index among fatty acid-samples, the complex index maximum increased as the chain length increased up to octanoic acid and then decreased. In comparison of complex index among fatty acid-, monoacylglycerol-, and polyglycerol fatty acid ester-samples at each acyl chain, the complex index maximum followed the order polyglycerol fatty acid ester > monoacylglycerol > fatty acid. Fatty acid, monoacylglycerol, and polyglycerol fatty acid ester with long acyl chains greatly reduced the enzymatic hydrolysis of starch. Polyglycerol fatty acid ester with palmitic acid chains was the strongest inhibitor of starch hydrolysis, suggesting that further complex formation may occur during the hydrolysis of gelatinized starch (enzyme-annealing).