abandon/ ə’bændən/ vt.丢弃；放弃，抛弃
aboard/ ə’bɔ:d/ ad.在船(车)上；上船
absolute/ ‘æbsəlu:t/ a.绝对的；纯粹的
absolutely/ ‘æbsəlu:tli/ ad.完全地；绝对地
absorb/ əb’sɔ:b/ vt.吸收；使专心
abstract/ ’æbstrækt/ n.摘要
abundant/ ə’bΛndənt/ a.丰富的；大量的
abuse/ ə’bju:z, ə’bju:s/ vt.滥用；虐待 n.滥用
academic/ ækə’demik/ a.学院的；学术的
accelerate/ æk’seləreit/ vt.(使)加快；促进
access/ ‘ækses/ n.接近；通道，入口
accidental/ æksi’dentl/ a.偶然的；非本质的
accommodate/ ə’kɔmədeit/ vt.容纳；供应，供给
accommodation/ ə,kɔmə’deiʃən/ n.招待设备；预定铺位
accompany/ ə’kΛmpəni/ vt.陪伴，陪同；伴随
accomplish/ ə’kɔmpliʃ/ vt.达到(目的)；完成
accordance/ ə’kɔr:dəns/ n.一致；和谐；授予
accordingly/ ə’kɔr:diŋli/ ad.因此，所以；照着
account/ ə’kaunt/ n.记述；解释；帐目
accumulate/ ə’kju:mjuleit/ vt.积累 vi.堆积
accuracy/ ‘ækjurəsi/ n.准确(性)；准确度
accurate/ ‘ækjurit/ a.准确的，正确无误的
accustomed/ ə’kΛstəmd/ a.惯常的；习惯的
acid/ ‘æsid/ n.酸；酸的，酸性的
acquaintance/ ə’kweintəns/ n.认识；了解；熟人
acquire / ə’kwaiə/ vt.取得；获得；学到
acre/ ‘eikə/ n.英亩(=6.07亩)
adapt/ ə’dæpt/ vt.使适应；改编
addition/ ə’diʃən/ n.加，加法；附加物
additional/ ə’diʃənl/ a.附加的，追加的
address / ə’dres/ n.地址；演说；谈吐
adequate/ ‘ædikwit/ a.足够的；可以胜任的
adjust/ ə’dʒΛst/ vt.调整，调节；校正
administration / ədminis’treiʃən/ n.管理；管理部门
admission/ əd’miʃən/ n.允许进入；承认
admit/ əd’mit/ vt.承认；准许…进入
advance/ əd’va:ns/ vi.前进；提高 n.进展
advanced/ əd’va:nst/ a.先进的；高级的
adventure/ əd’ventʃə/ n.冒险；惊险活动
advisable/ əd’vaizəbl/ n.明智的；可取的
affair/ ə’feə/ n.事情，事件；事务
affect/ ə’fekt/ vt.影响；感动
affection/ ə’fekʃən/ n.慈爱，爱；爱慕
afford/ ə’fɔr:d/ vt.担负得起…；提供
afterward/ ‘a:ftəwəd(z)/ ad.后来，以后
age/ eidʒ/ vt.变老
aggressive/ ə’gresiv/ a.侵略的；好斗的
aircraft/ ‘eəkra:ft/ n.飞机，飞行器
alarm/ ə’la:m/ n.惊恐，忧虑；警报
alcohol/ ‘ælkəhɔl/ n.酒精，乙醇
alike/ ə’laik/ a.同样的，相同的
alloy/ ‘ælɔi, ə’lɔi/ n.合金；(金属的)成色
alphabet/ ‘ælfəbit/ n.字母表，字母系统
alter/ ‘ɔ:ltə/ vt.改变，变更；改做
alternative/ ɔ:l’tə:nətiv/ n.替换物；取舍，抉择
altitude/ ‘æltitju:d/ n.高，高度；高处
aluminium/ ælju’minjəm/ n.铝
amaze/ ə’meiz/ vt.使惊奇，使惊愕
ambulance/ ‘æmbjuləns/ n.救护车；野战医院
amongst/ ə’mΛŋst/ prep在…之中(=among)
amuse/ ə’mju:z/ vt.逗…乐；给…娱乐
analyse/ ‘ænəlaiz/ vt.分析，分解，解析
analysis/ ə’næləsis/ n.分析，分解，解析
ancestor/ ‘ænsistə/ n.祖宗，祖先
anchor/ ‘æŋkə/ n.锚 vi.抛锚，停泊
ancient/ ‘einʃənt/ a.古代的，古老的
ankle/ ‘æŋkl/ n.踝，踝节部
announce/ ə’nauns/ vt.宣布，宣告，发表
annoy/ ə’nɔi/ vt.使恼怒；打搅
annual/ ‘ænjuəl/ a.每年的 n.年报
anticipate/ æn’tisipeit/ vt.预料，预期，期望
anxiety/ æŋg’zaiəti/ n.焦虑，忧虑；渴望
anxious/ ‘æŋkʃəs/ a.忧虑的；渴望的
apart/ ə’pa:t/ ad.相隔；分开；除去
apologize/ ə’pɔlədʒaiz/ vi.道歉，谢罪，认错
apparatus/ ,æpə’reitəs/ n.器械，仪器；器官
appeal/ ə’pi:l/ vi.&n.呼吁；申述
appetite/ ‘æpitait/ n.食欲，胃口；欲望
appliance/ ə’plaiəns/ n.用具，器具，器械
applicable/ ‘æplikəbl/ a.能应用的；适当的
application/ æpli’keiʃən/ n.请求，申请；施用
appɔint/ ə’pɔint/ vt.任命，委任；约定
appreciate/ ə’pri:ʃieit/ vt.欣赏；领会；感谢
approval/ ə’pru:vəl/ n.赞成，同意；批准
approve/ ə’pru:v/ vt.赞成，称许；批准
approximate/ ə’prɔksimit/ a.近似的 vt.近似
arbitrary/ ‘a:bitrəri/ a.随心所欲的；专断的
architecture/ ‘a:kitektʃə/ n.建筑学；建筑式样
argue/ ‘a:gju:/ vi.争论，争辩，辩论
argument/ ‘a:gju:mənt/ n.争论，辩论；理由
arise/ ə’raiz/ vi.出现；由…引起
arithmetic/ ə’riθmətik/ n.算术，四则运算
arouse/ ə’rauz/ vt.引起，唤起；唤醒
article/ ‘a:tikl/ n.条款；物品
artificial/ a:ti’fiʃəl/ a.人工的；娇揉造作的
artistic/ a:’tistik/ a.艺术的；艺术家的
ash/ æʃ/ n.灰，灰末；骨灰
ashamed/ ə’ʃeimd/ a.惭愧(的)；羞耻(的)
aspect/ ‘æspekt/ n.方面；样子，外表
assemble/ ə’sembl/ vt.集合，召集；装配
assembly/ ə’sembli/ n.集合；集会；装配
assess/ ə’ses/ vt.对(财产等)估价
assign/ ə’sain/ vt.指派；分配；指定
assist/ ə’sist/ vt.援助，帮助；搀扶
assistance/ ə’sistəns/ n. 协助,援助
associate/ ə’səuʃieit/ vi.交往 n.伙伴，同事
association/ əsəusi’eiʃən/ n.协会，团体；联合
assume/ ə’sju:m/ vt.假定；承担；呈现
assure/ ə’ʃuə/ vt.使确信；向…保证
astonish/ əs’tɔniʃ/ vt.使惊讶，使吃惊
astronaut/ ‘æstʃəunɔ:t/ n.宇宙航行员，宇航员
ætlantic/ ət’læntik/ a.大西洋的 n.大西洋
atom/ ‘ætəm/ n.原子；微粒；微量
atomic/ ə’tɔmik/ a.原子的；原子能的
attach/ ə’tætʃ/ vt.缚，系，贴；附加
attain/’tein/ vt.达到，获得，完成
attempt/ ə’tempt/ vt.尝试，试图 n.企图
attend/ ə’tend/ vt.出席；照顾，护理
attribute/ ‘ætribju:t/ vt.把…归因于 n.属性
audience/ ‘ɔ:djəns/ n.听众，观众，读者
authority/ ɔ:’θɔriti/ n.当局，官方；权力
automatic/ ɔ:tə’mætik/ a.自动的；机械的
automobile/ ‘ɔ:təməbi:l/ n.汽车，机动车
auxiliary/ ɔ:g’ziljəri/ a.辅助的；附属的
available/ ə’veiləbl/ a.可利用的；通用的
avenue/ ‘ævinju:/ n.林荫道，道路；大街
await/ ə’weit/ vt.等候，期待
awake/ ə’weik/ a.醒着的 vt.唤醒
award/ ə’wɔ:d/ n.奖，奖品；判定
aware/ ə’weə/ a.知道的，意识到的
awful/ ‘ɔ:ful/ a.令人不愉快的
awkward/ ‘ɔ:kwəd/ a.笨拙的；尴尬的
ax/ æks/ n.斧子
baby/ ‘beibi/ n.婴儿；孩子气的人
back/ bæk/ ad.在后；回原处；回
background/ ‘bækgraund/ n.背景，后景，经历
backward/ ‘bækwəd/ a.向后的；倒的 ad.倒
bacteria/ bæk’tiəriə/ n.细菌
bad/ bæd/ a.坏的，恶的；严重的
badly/ ‘bædli/ ad.坏，差；严重地
bag/ bæg/ n.袋，包，钱包，背包
baggage/ ‘bægidʒ/ n.行李
bake/ beik/ vt.烤，烘，焙；烧硬
balance/ ‘bæləns/ vt.使平衡；称 n.天平
ball/ bɔ:l/ n.球，球状物；舞会
balloon/ bə’lu:n/ n.气球，玩具气球
banana/ bə’na:nə/ n.香蕉；芭蕉属植物
band/ bænd/ n.乐队；带；波段
bang/ bæŋ/ n.巨响，枪声；猛击
bank/ bæŋk/ n.银行；库；岩，堤
bar/ ba:/ n.酒吧间；条，杆；栅
barber/ ‘ba:bə/ n.理发师
bare/ beə/ a.赤裸的；仅仅的
bargain/ ‘ba:gin/ n.交易 vi.议价；成交
barrel/ ‘bærəl/ n.桶；圆筒；枪管
barrier/ ‘bæriə/ n.栅栏，屏障；障碍
base/ beis/ n.基础，底层；基地
basic/ ‘beisik/ a.基本的，基础的
basically/ ‘beisikəli/ ad.基本上
basin/ ‘beisn/ n.盆，洗脸盆；盆地
basis/ ‘beisis/ n.基础，根据
basket/ ‘ba:skit/ n.篮，篓，筐
basketball/ ‘ba:skitbɔ:l/ n.篮球；篮球运动
bath/ ba:θ/ n.浴，洗澡；浴缸
bathe/ beið/ vt.给…洗澡；弄湿
bathroom/ ‘ba:θrum/ n.浴室；盥洗室
battery/ ‘bætəri/ n.电池；一套，一组
battle/ ‘bætl/ n.战役；斗争 vi.作战
bay/ bei/ n.湾；山脉中的凹处
be/ bi:/ aux.v.&vi.是，在，做
beach/ bi:tʃ/ n.海滩，湖滩，河滩
beam/ bi:m/ n.梁；横梁；束，柱
bean/ bi:n/ n.豆，蚕豆
bear/ beə/ n.熊；粗鲁的人
bear/ beə/ vt.容忍；负担；生育
beard/ biəd/ n.胡须，络腮胡子
beast/ bi:st/ n.兽，野兽；牲畜
beat/ bi:t/ vt.&vi.打，敲；打败
beautiful/ ‘bju:tiful/ a.美的，美丽的
beauty/ ‘bju:ti/ n.美，美丽；美人
because/ bi’kɔz/ conj.由于，因为
become/ bi’kΛm/ vi.变成；成为，变得
bed/ bed/ n.床，床位；圃；河床
bee/ bi:/ n.蜂，密蜂；忙碌的人
beef/ bi:f/ n.牛肉；菜牛
beer/ biə/ n.啤酒
before/ bi’fɔ:/ prep.在…以前；向…
beg/ beg/ vt.&vi.乞求；请求
begin/ bi’gin/ vi.开始 vt.开始
beginning/ bi’giniŋ/ n.开始，开端；起源
behalf/ bi’ha:f/ n.利益，维护，支持
behave/ bi’heiv/ vi.表现，举止；运转
behavior/ bi’heivjə/ n.行为，举止，态度
behind/ bi’haind/ prep.在…后面
being/ ‘bi:iŋ/ n.存在；生物；生命
belief/ bi’li:f/ n.信任，相信；信念
believe/ bi’li:v/ vt.相信；认为
bell/ bel/ n.钟，铃，门铃；钟声
belong/ bi’lɔŋ/ vi.属于，附属
below/ bi’ləu/ prep.在…下面(以下)
belt/ belt/ n.带，腰带；皮带；区
bench/ bentʃ/ n.长凳，条凳；工作台
bend/ bend/ vt.使弯曲 vi.弯曲
beneath/ bi’ni:θ/ prep.在…下方
beneficial/ beni’fiʃəl/ a.有利的，有益的
benefit/ ‘benifit/ n.利益；恩惠；津贴
beside/ bi’said/ prep.在…旁边
besides/ bi’saidz/ ad.而且prep.除…之外
best/ best/ a.最好的；最大的
bet/ bet/ vt.&vi.&n.打赌
better/ ‘betə/ a.较好的 ad.更好地
between/ bi’twi:n/ prep.在…中间
beyond/ bi’jɔnd/ prep.在…的那边
Bible/ ‘baibl/ n.基督教《圣经》
bicycle/ ‘baisikl/ n.自行车，脚踏车
big/ big/ a.大的，巨大的
bike/ baik/ n.自行车 vi.骑自行车
bill/ bil/ n.账单；招贴；票据
billion/ ‘biljən/ num.万亿(英)
bind/ baind/ vt.捆绑；包扎；装钉
biology/ bai’ɔlədʒi/ n.生物学；生态学
bird/ bə:d/ n.鸟，禽
birth/ bə:θ/ n.分娩，出生；出身
birthday/ ‘bə:θdi/ n.生日，诞生的日期
biscuit/ ‘biskit/ n.(英)饼干；(美)软饼
bit/ bit/ n.一点，一些，小片
bite/ bait/ vt.咬，叮，螫；剌穿
bitter/ ‘bitə/ a.痛苦的；严寒的
black/ blæk/ a.黑色的；黑暗的
blackboard/ ‘blækbɔ:d/ n.黑板
blade/ bleid/ n.刀刃，刀片；叶片
blame/ bleim/ vt.责备，把…归咎于
blank/ blæŋk/ a.空白的 n.空白
blanket/ ‘blæŋkit/ n.毛毯，毯子，羊毛毯
blast/ bla:st/ n.爆炸，冲击波 vt.炸
bleed/ bli:d/ vi.出血，流血；泌脂
blend/ blend/ vt.&vi.&n.混和
blind/ blaind/ a.瞎的；盲目的
block/ blɔk/ n.街区 vt.堵塞，拦阻
blood/ blΛd/ n.血，血液；血统
bloom/ blu:m/ n.花；开花，开花期
blow/ bləu/ vi.吹，吹动；吹响
blue/ blu:/ a.蓝色的 n.蓝色
board/ bɔ:d/ n.板 vt.上(船、车等)
boast/ bəust/ vi.自夸 vt.吹嘘
boat/ bəut/ n.小船，艇；渔船
body/ ‘bɔdi/ n.身体；主体；尸体
bɔil/ bɔil/ vi.沸腾；汽化vt.煮沸
bold/ bəuld/ a.大胆的；冒失的
bolt/ bəult/ n.螺栓；插销 vt.闩门
bomb/ bɔm/ n.炸弹 vt.轰炸
bond/ bɔnd/ n.联结，联系；公债
bone/ bəun/ n.骨，骨骼
book/ buk/ n.书，书籍 vt.预定
boot/ bu:t/ n.靴子，长统靴
border/ ‘bɔ:də/ n.边，边缘；边界
bore/ bɔ:/ vt.使厌烦；钻，挖
born/ bɔ:n/ a.天生的；出生的
borrow/ ‘bɔrəu/ vt.借，借用，借人
boss/ bɔs/ n.老板，上司 vt.指挥
both/ bəuθ/ pron.两者(都)
bother/ ‘bɔðə/ vt.烦扰，迷惑 n.麻烦
bottle/ ‘bɔtl/ n.瓶，酒瓶；一瓶
bottom/ ‘bɔtəm/ n.底，底部，根基
bounce/ bauns/ vi.反跳，弹起；跳起
bound/ baund/ a.一定的；有义务的
boundary/ ‘baundəri/ n.分界线，办界
bow/ bau/ n.弓；蝴蝶结；鞠躬
bowl/ bəul/ n.碗，钵；碗状物
box/ bɔks/ n.箱，盒；包箱
box/ bɔks/ vi. 拳击，打拳
boy/ bɔi/ n.男孩，少年；家伙
brain/ brein/ n.脑，脑髓；脑力
brake/ breik/ n.闸，刹车 vi.制动
branch/ bra:ntʃ/ n.树枝；分部；分科
brand/ brænd/ n.商品；烙印 vt.铭刻
P(A | B) is a conditional probability: the likelihood of event A occurring given that B is true.
P(B | A) is also a conditional probability: the likelihood of event B occurring given that A is true.
P(A) and P(B) are the probabilities of observing A and B independently of each otherthis is known as the marginal probability.
Bayes theorem Interpretations:
Bayesian inference derives the posterior probability as a consequence of two antecedents: a prior probability and a " likelihood function " derived from a statistical model for the observed data. Bayesian inference computes the posterior probability according to Bayes' theorem:
H: stands for any hypothesis whose probability may be affected by data (called evidence below). Often there are competing hypotheses, and the task is to determine which is the most probable.
P(H): the prior probability , is the estimate of the probability of the hypothesis H before the data E, the current evidence, is observed.
P(H | E): the posterior probability , is the probability of H given E, i.e., after E is observed. This is what we want to know: the probability of a hypothesis given the observed evidence.
P(E | H): is the probability of observing E given H, and is called the likelihood . As a function of E with H fixed, it indicates the compatibility of the evidence with the given hypothesis. The likelihood function is a function of the evidence, E, while the posterior probability is a function of the hypothesis, H.
P(E): is sometimes termed the marginal likelihood or " model evidence ". This factor is the same for all possible hypotheses being considered (as is evident from the fact that the hypothesis H does not appear anywhere in the symbol, unlike for all the other factors), so this factor does not enter into determining the relative probabilities of different hypotheses.
Sometimes, Bayes theorem can be written as:
where the factor P(E | H) / P(E) can be interpreted as the impact of E on the probability of H .
Binomial distribution with parameters n and p is the discrete probability distribution of the number of successes in a sequence of n independent experiments , each asking a yes–no question , and each with its own Boolean-valued outcome: a random variable containing a single bit of information: success/yes/true/one (with probability p) or failure/no/false/zero (with probability q = 1 − p).
In general, if the random variable X follows the binomial distribution with parameters n ∈ ℕ and p ∈ [0,1], we write X ~ B(n, p) . The probability of getting exactly k successes in n trials is given by the probability mass function :
The cumulative distribution function can be expressed as:
Mean: E(X) = npVariance: Var(X) = npq = np(1-q)Mode:
If two binomially distributed random variables X and Y are observed together, estimating their covariance can be useful. The covariance is Cov(X,Y) = E(XY) - μX * μY
In the case n = 1 (the case of Bernoulli trials ) XY is non-zero only when both X and Y are one, and μ X and μ Y are equal to the two probabilities. Defining p B as the probability of both happening at the same time, this gives
In a bivariate setting involving random variables X and Y, there is a particular expectation that is often of interest. It is called covariance and is given by: Cov(X,Y) = E((X-E(X))(Y-E(Y)) where the expectation is taken over the bivariate distribution of X and Y. Alternatively, Cov(X,Y) = E(XY) - E(X)E(Y)
Moreover, a scaled version of covariance is the correlation ρ which is given by
ρ = Corr(X,Y) = Cov(x,y) / [sqrt(Var(X)*sqrt(Var(Y)], Var(X)=σx^2
Assume that total number of successes X ~ B(n,p) with np>=5, n(1-p)>=5 so that the normal approximation to the binomial is reasonable.
In practice, p is unknown. Under the normal approximation, we have X ~ N(np, np(1-p)) and we define p^ = X/n as the proportion of successes. Since p^ is a linear combination of normal random variable, it follows that p^ ~ N(p,p(1-p)/n) then the probability statement is
Let Za/2 denote the (1-a/2)100-th percentile for the standard normal distribution, a (1-a)100% approximation confidence interval ( because we user normal distribution to the binomial and the substitution of p with p-hat ) for p-hat is given by
the normal (or Gaussian) distribution is a very common continuous probability distribution. Normal distributions are important in statistics and are often used in the natural and social sciences to represent real-valued random variables whose distributions are not known.
If X ~ N(μ,σ^2), then E(X) = μ, and Var(X) = σ^2
σ^2 is the variance and not the standard deviation.
A random variable Z ~ N(0,1) is referred to as standard normal and it has the simplified pdf:
The relationship between an arbitrary normal random variable X ~ N(μ, σ^2) and the standard normal distribution is expressed via (X-μ) / σ ~ N(0,1)
In this case, we assume X1,X2,...,XN iid normal(μ, σ^2) where our interest concerns that μ is unknown and σ is known for ease of development. (In real world, we can't find a case with known σ &unknown μ)
X-bar ~ N(μ, σ^2/n)
Rearranging terms:
Finally, we obtain a 95% confidence interval (as follows) for μ
More generally, let Za/2 denote the (1-a/2)100-th percentile for the standard normal distribution, a (1-a)100% confidence interval for μ is given by
we use the observed value x-bar. It is understood that confidence intervals are functions of observed statistics.
It concerns the presentation of data (either numerical or graphical) in a way that makes it easier to digest data.
outliers : too big or small
centrality : values in the middle portion of the dotplot
dispersion : spread or variation in the data
modality : a histogram with two distinct humps is referred to as bimodal
skewness :
symmetry :
How to choose interval as x-axis: choose the number of intervals roughly equal to sqrt(n) where n is the number of observations .
For those intervals are not equal length, we should plot relative frequency divided by intervals length on the vertical axis, instead of using frequency .
sample median ( Q2 )top-edge is 3/4 quantile ( Q3 )bottom-edge: 1/4 quantile ( Q1 )
interquartile range ( IQR ) : Q3-Q1, known as ΔQ
maximum interval： Q3+1.5ΔQ or 90th percentile
minimum interval： Q1-1.5ΔQ or 10th percentile
values that out of max &min intervals are Outliers .
whiskers (vertical dashed lines) extend to the outer limits of the data and circles correspond to outliers.
extrapolated data : when predicting, you should be cautious about predictions based on extrapolated data. There perhaps appears a positive increase trend from the pairplot with two variables X,Y, but it doesn't mean they have the same relationship for X, Y. (Data should be combined with the real world)
It is a numerical descriptive statistic for investigating paired data is the sample correlation or correlation or correlation coefficient r defined by
Properties:
-1 <= r <= 1
when r close to 1, the points are clustered about a line with positive slope
when r close to -1, the points are clustered about a line with negative slope
when r close to 0, points are lack of linear relationship . However, there may be a quadratic relationship
when x and y are correlated (not close to 0), it merely denotes the presence of a linear association. For example, weight and height are positively correlated, and it is obviously wrong to state that one causes the other.
In order to establish cause and effect relationship, we should do a controlled study .
the average of the results obtained from a large number of trials should be close to the expected value, and will tend to become closer as more trials are performed.
For the CLT, we assume that the random variables X1,X2,...,Xn are * iid from a population with mean μ and variance σ^2. The CLT states that as n =>infinity, the distribution of * (X_bar - μ)/(σ/sqrt(n)) converges to the distribution of a standard normal random variable.
从一个均值为 μ 、标准差为σ的总体中选取一个有n个观测值的随机样本。那么当n足够大时， x¯的抽样分布将近似服从均值μx¯=μ、标准差σx¯=σ/√n的正态分布 。并且样本量越大，对x¯的抽样分布的正太近似越好
In probability theory , the central limit theorem ( CLT ) establishes that, in some situations, when independent random variables are added, their properly normalized sum tends toward a normal distribution (informally a " bell curve ") even if the original variables themselves are not normally distributed.
要求：
1. 总体本身的分布不要求正态分布
2. 样本每组要足够大，但也不需要太大 n≥30
中心极限定理在理论上保证了我们可以用只抽样一部分的方法，达到推测研究对象统计参数的目的。
linear regression is predicting the value of a variable Y(dependent variable) based on some variable X(independent variable) provided there is a linear relationship between X and Y.
Y=b0 + b1X+e
(Recall that the regression equation without the error term, Y=b0 + b1X , is called the least squares line .)
SSTO, a.k.a SST , sum of squared total: sum of difference from the mean of y and data point yi
SSE , sum of squared error: sum of difference from the estimated regression line and data point yi
SSR , sum of squared regression: quantifies how far the estimated sloped regression line , y^i, is from the horizontal " no relationship line ," the sample mean or y¯.
From the above example, it tells us that most of the variation in the response y ( SSTO = 1827.6) is just due to random variation ( SSE = 1708.5), not due to the regression of y on x ( SSR = 119.1).
If r^ 2 = 1, all data points fall perfectly on the regression line. The predictor x accounts for all of the variation in y !
If r^ 2 = 0, the estimated regression line is perfectly horizontal. The predictor x accounts for none of the variation in y !
r^ 2 ×100 percent of the variation in y is 'explained by' the variation in predictor x .
SSE is the amount of variation that is left unexplained by the model.
1. The coefficient of determination r^ 2 and the correlation coefficient r quantify the strength of a linear relationship . It is possible that r^ 2 = 0% and r = 0, suggesting there is no linear relation between x and y , and yet a perfect curved (or "curvilinear" relationship ) exists.
[Most misinterpreting concept] 2. A large r^2 value should not be interpreted as meaning that the estimated regression line fits the data well .
Although the R-squared value is 92% and only 8% of the variation US population is left to explain after taking into account the year in a linear way. The plot suggests that a curve plot describe the relationship even better. (Its large value does suggest that taking into account year is better than not doing so . It just doesn't tell us that we could still do better .)
3. The coefficient of determination r2 and the correlation coefficient r can both be greatly affected by just one data point (or a few data points) .
4. Correlation (or association) does not imply causation .
VIF check the co-linearity between explanatory variables. Over 5 is too bad.
H0: null hypothesisH1: alternative hypothesis.
Testing begins by assuming that H0 is true, and data is collected in an attempt to establish the truth of H1.
H0 is usually what you would typically expect (ie, H0 represents the status quo).
In inference step, we calculate a p-value, defined as the probability of observing data as extreme or more extreme (in the direction of H1) than what we observed given that H0 is true.
Significance level: a, usually equal to 0.01, 0.05
If p-value is less than a, reject H0
If p-value is larger than a, fail to reject H0.
......
When fitting models, it is possible to increase the likelihood by adding parameters, but doing so may result in overfitting . Both BIC and AIC attempt to resolve this problem by introducing a penalty term for the number of parameters in the model.
AIC Akaike information criterion: 2k - 2ln(L) where k is the number of parameters in the model (or the number of degrees of freedom being used up)ln(L) is the ' log likelihood ', which is a measure of how well the model fits the data. Low AIC is better. 2k is the 'penalty' term.
AIC measure the Goodness of fit &Complexity (number of terms)
Comparing AIC with the proportion of variance explained, R^2, R^2 only measures goodness of fit.
However, because of co-linearity, sometimes that variable is 'stealing' the significance from some other term. The AIC doesn't care which terms are significant , it just looks at how well the model fits as a whole.
BIC Bayesian Information Criterion: (ln(n)*k) - 2ln(L) where n is the number of observations, also call the sample size, k stands for the number of parameters (df).
BIC is similar to the AIC, but imposes a larger penalty term for complexity . Lower BIC is better. And BIC favors for simpler models, given a set of candidate models . What's more, BIC is easier to find significance in variables that are unimportant when n is large because of large penalty.
we also need to check influential outliers, homoscedasticity (equal variance) and normality.
Residual is to check above mentioned properties.
To check normality: use Shapiro-Wilks Test
It is a hypothesis test whose null hypothesis is ' your data is normally distributed '
Large p-value, fail to reject H0, you have no evidence against normalitysmall p-value, reject H0, so you have evidence of non-normality
To check homoscedasticity: use Levene Test
Still hypothesis with null hypothesis: all input samples are from populations with equal variances .
Outlier Detection: in statistical method, not mention approaches in data mining aspect.
noise: it is random error or variance in a measured variable
noise should be removed before outlier detection.
outlier: A data object that deviates significantly from the normal objects as if it were generated by a different mechanism . It violates the mechanism that generates the normal data .
Parametric Methods I : detection univariate outliers based on Normal Distribution
μ+3σ region contains 99.7% data, outliers are out of this region .
Parametric Methods II : detection of multivariate outliers.
bottom line: transform the multivariate outlier detection task into a univariate outlier detection problem
use X^2-statistic: (chi square statistic)
If X^2-statistic is large, then Object Oi is an outlier.
A low value for chi-square means there is a high correlation between your two sets of data. In theory, if your observed and expected values were equal (“no difference”) then chi-square would be zero — an event that is unlikely to happen in real life. You could take your calculated chi-square value and compare it to a critical value from a chi-square table. If the chi-square value is more than the critical value, then there is a significant difference.
[ Omit ] Parametric Methods III: Using mixture of parametric distributions
Outlier Detection is a big topic that can be expand for another article. Let me stop it here in Statistics topic.
Note that statistics are quantities that we can calculate , and therefore do not depend on unknown parameters . Moveover, statistics have associated probability distributions, and we are sometimes interested in the distributions of statistics.
MLE： maximum likelihood estimate 最大似然估计
MSE: mean squared error 误差均方
RMSE: root mean squared error 误差均方根
r^2: coefficient of determination 确定系数
SE: standard error 标准误
SEM: standard error of the mean 均数的标准误
SS: sum of squares 平方和
SSE: sum of squared error of the prediction function
SSR: sum of squared residuals
SST: total sum of squares
1、东南亚木材（12种）1.1 学名：印茄intsia biujga 0.Ktze。
商品名：梅宝Merbau（马来西亚、印度尼西亚）； Ipil（菲律宾）； Kwila（巴布亚新儿内亚）； Hintzy（马达加斯加）。
俗称：菠萝格、南洋木宝
木材材性：边格宽4-8cm，淡白色，与心边区别明显，生材的心材浅黄色，在外界作用下变成暗褐色；浅色的薄壁组织在旋切面上常构成有装饰作用的花纹；木材具有光泽；有树脂气味；纹理直，有时呈混乱的波纹状；结构粗而均匀；干缩小，干缩率生材至气干材径向0.9-3.1%，弦向1.6-4.1%；木材重而硬，气干密度约0.9g/cm3，强度甚高；木材的耐腐性强。
用途：由于木材重而硬，且强度高，并且具有一定的花纹，所以多用于要求木材耐久、强度大和有装饰性的场合，如 ：建筑构件、高级家具、细木工、地板等。
1.2 学名：平滑（重黄）娑罗双 Shorea Laevis Ridl.
商品名：巴劳Balau（马来西亚）； Selangan batu kumus（沙巴、沙捞越）； AK、Teng、 Ack（泰国）； Bangkirai（印度尼西亚）； Yakal、Malagkal、Guiuo（菲律宾）。
俗称：黄梢、油抄、金柚檀。
木材材性：心材黄褐色，新伐时黄或灰褐色或带红，与边材区别略明显，边材色浅，宽3-6cm；生长轮不明显，有时介以不明显的纤维；木材光泽差；无特殊气味；纹理深交错；结构细且均匀；干缩率生材至气干径向1.8%，弦向3.7%；木材重且质地硬，马来西亚产的该树种气干密度约0.96g/cm3。侧面硬度为10010N，强度甚高；木材的耐腐性很强。
用途：用于重要的建筑物构件，桥梁、枕木、电杆、造船、承重地板、承重家具、细木工和门窗框等；由于木材耐腐性强，适用于条件差的场合。
1.3 学名：疏花（深红）娑罗双Shorea pauciflora King.
商品名：红柳安 Red lauan、Tangile、Tiaong（菲律宾）；深红麻兰蒂 Dark red maranti、Nemusu（马来西亚）； 0bar Suluk（沙巴）； Meranti merah、Meranti Ketuko（印度尼西亚）。
木材材性：心材红至深红褐色，边材桃红色，心边材区别略明显；生长轮不明显；木材光泽弱；无特殊气味；纹理交错；结构略粗且均匀；干缩率生材到炉干径向2.2%，弦向0.7；木材略耐腐，边材易被粉蠹虫和白蚁危害，不抗海生钻木动 物侵袭；木材重量中等，马来西亚的该树种的气干密度约0.68g/cm3木材强度低于至中。
1.4 学名：角香茶茱萸 Cantleyt corniculata（Becc） Howard
商品名：达茹-达茹Daru-Daru；德达茹Dedaru（马来西亚）； Seranai（印度尼西亚）； Bedaru（沙捞越、印度尼西亚）； Samala（沙巴）。
俗称：芸香
木材材性：心材黄褐色，与边材区别略明显，边材浅黄褐色；生长轮不明显；木材具有汹涌新切面具有香味；耐腐性甚强；纹理交错；结构细而均匀；木材理，马来西亚产的该树种的气干密度约0.9g/cm3；强度高至甚高。
用途：可用于重型建筑的构件、桥梁、车辆、地板、工具等，适宜各种需要强度大和耐久的地方。
1.5 学名：木荚豆Xylia xylocarpa（Roxb.）Taub
商品名：卡姆-贼Cam-xe（柬埔寨、泰国、越南）；Pyinkado（缅甸）； Irul（印度尼西亚）； Sokram（柬埔寨）；Deng（泰国）。
俗称：金车木、红荚木。
木材材性：心材红褐色，具有深钩的带状条纹，与边格区别明显，边材窄且浅红白色；生长轮明显；由于有树脂沉淀，木材带深色的斑点；木材具有光泽；无特殊气味；纹理不规则交错；结构细且均匀；体积干缩率11-12%；耐腐性很强；木材甚重，平均气干密度约为0.99g/cm3；强度大。
用途：重要建筑物的构件、桥梁、地板、装饰板、工具、桅杆、矿柱等，比较适用于混度大的场所。
1.6 学名：马来甘巴豆 Koompassia malaccensis Maing
商品名：康派斯、克姆帕斯Kempas；门格拉斯Mengeris（加里曼丹岛）； Empas（沙巴）；Impas（婆罗洲、印度尼西亚、沙巴）；Taulong（马来西亚）； Upil（印度尼西亚）； Bueng（泰国）。
俗称：金不换、南洋钢柏木。
木材材性；边材与心材有明显的区别，边材呈白或浅黄色，大径木的边材宽5cm，新切口心材呈褐红色，在空气的作用下变成桔红色，并带黄褐色的细线条；木材具有光泽；无特殊气味；耐腐性强，但容易受白蚁和粉蠹虫的危害；纹理交替，有的呈波浪状；结构粗而均匀；干缩小，干缩率生材至气干径向2%，弦向3%；木材重至甚重且硬，印度尼西亚的该树种的气干密度可达1.06g/cm3强度高，特别是抗压强度。
用途：可用重要的建筑和构筑物：枕木、车辆、地板、家具、工具柄、用于化学工业上的木桶和木槽、由于该木材有轻度的酸性反应，因此不宜使其与黑色金属接触。
1.7 学名：柚木 Tectona grandis Linn
商品名：柚木Teak（缅甸、印度尼西亚、Jati（印度尼西亚）；Kyun（缅甸）； Maisak（泰国）。
木材材性：心材黄褐色，褐色，久则呈暗褐色，有时带深条状，边材浅黄或浅白色，必材和边材区别明显；年轮清晰；木材具有光泽且表面油腻感：无特殊气味；纹理下略交错；结构粗且不均交；干缩小，干缩率从生材至气干径向2.2%，弦向4.0%；很耐腐，能抗海生钻木动物危害；木材重量中等，缅甸柚木的气干密度0.64g/cm3左右；强度低至中。
1.8 学名：大花龙脑香Dipterocarpus grandiflorus Bianco
商品名：克隆Keruing；阿必通Apitong、Hagakhak（菲律宾）；Gurjun（印度）；Keruing belimbing（马来西亚、北波罗洲）： Kanyinbyan（缅甸）。
俗称：南洋油崽木
木材材性；心材灰红褐色至红褐色，并在空气作用下变深，边材浅灰褐色，与心材区别略明显；木材光泽弱，常有树脂气味；纹理通直；结构略粗且略均匀；木材的天然缺陷很少；较耐腐；干缩甚大，干缩率生材至炉干的径向缩率为7.0%，弦向为12.9%；木材硬度较大，重量较重，马来西亚的该权势种的气干密度0.80g/cm3左右；木材强度较大，特别是耐载荷冲击的强度。
用途：主要用于轻型构件、重要建筑物和构筑物、地板等；由于木材的抗酸和抗化学药剂性能可以用作试验室的装修及内部器具。
1.9 学名：坤甸铁木 Eusideroxylon xwageri Teijsm＆Binnend
商品名：贝联Beilian（沙巴、沙捞越、印度尼西亚）；Tambulian（沙巴、菲律宾）；Bormeo iromwood（欧洲）；0nglen；Ulin等。
俗称：铁梨木、铁木、铁檀。
木材材性；心材黄褐色至红褐色，久置于大气中转呈黑色，边材金黄金，心格与边材区别明显；木材具有光泽；新切面有柠檬味；木材纹理通直或稍有交错；结构细且均匀；木材耐腐性强，但易受海生动物危害；木材干缩甚大，干缩率生材至炉干径向4.2%，弦向8.3%；木材甚重且硬，印度尼西亚的该树种气干密度1.19g/cm3左右；心材强度甚高。
用途：因为木材重、硬、强度大又耐久。所以可以作为建筑构件、桥梁、电杆、地板、家具等。
1.10 学名：阔叶黄檀、印度玫瑰木Dalbergia latifolia Roxb
商品名：玫瑰木Rosewood（印度、新加坡、缅甸）；Indian Rosewood Bombayblack-wood （印度）；Sonkeling、Angsana Keling、Sonobrits、Java-palisandre（印度尼西亚）。
俗称：印尼些檀
木材材性：心材材色变异很大，以金黄褐色到深紫色，并带有深色条纹，时间久可能变成黑色，边材浅黄白色，常带有紫色窄条纹；木材具有光泽；咯有香味；结构细而均匀；纹理交错；很耐腐；干缩较大，干缩率轻和2.9%，弦向6.4%；材质重且硬；气干密度0.87g/cm3左右（是印度材质最硬的树种之一）；木材的刚度、抗剪和抗压强度、载荷冲击强度较
大。
用途：由于木材强度大又耐腐，而且材色和花纹漂亮，所以可以做家具、地板、装饰胶合板。
1.11 学名：阔萼摘亚木 Dialium platysepalum Baker
商品名：克然吉Keranji（沙巴、印度尼西亚）；Kerandjiasap（印度尼西亚）； Keranji Kuning besar（马来西亚）； Yi thong bueng（泰国）。
俗称：南洋红檀。
木材材性：心材新伐时为金黄褐色，久之则转深，为褐色和深红褐色，边材新伐时为白色，久在大气中变为浅褐色，心边材区别明显；木材光泽强，在径切面在有带状条纹，在弦切面上有之字型的图案；木材纹理呈波浪状或扭曲；结构一般且均匀；较耐腐；干缩小，干缩率生材至气干径向2.3%，弦向3.7%；材质较且较硬，气干密度约0.93-1.08g/cm3木材强度很高。
用途：可用于房屋建筑及室内时候、装饰单板、造船、家具及各种工农具柄。
1.12 学名：番龙眼、水黄皮pometia pinnata Forst
商品名：麻芦盖Malugay、Agupanga（菲律宾）；卡赛Kasai（东南亚、索罗门群岛）；唐Taun（巴布亚新几内亚）；Truong（越南）；Lan doeng、Kasi besar daun、Matoa（印度尼西亚）；Sibu（沙巴）。
俗称：缅甸红。
木材材性：心材红褐色或紫红褐色，边材浅红褐色，通常心边材区别不明显；生长轮略明显；木材具有光泽；无特殊气味；木材纹理直至略交错；结构甸而均匀；干缩大，干缩率从生材至含水率12%时的径向为3.1%，弦向为6.1%；木材略耐腐，易感染小蠹虫及海生钻木动物危害；木材略重且坚硬，马来西亚的该树种的气干密度约为0.74g/cm3；木材的强度中等。
用途：建筑上用的构件、地板、室内装饰等。2 非洲木材（12种）
2.1 学名：安哥拉紫檀Pterocarpus angolensis D.C
商品名：穆尼加MunigaGirassonde（安哥拉）；Ambila（莫桑比克）； Mukwa、 Muninga （赞比亚、津巴布韦）Kiaat、 Kajat、 Kajaatenhout（南非）； Mninga（坦桑尼亚）。木材材性：木材系半环孔至散孔材；边材浅灰或黄色，宽度3-5cm，心材材色变异大，从褐色到紫褐公，有时具有深色条纹，与边材区别明显；生长轮略明显：木材有光泽，有微弱香气；纹理直至略交错；结构细略均匀略耐腐，抗蚁和抗海生钻木动物能力较强；干缩小；木材重量中等，气干密度约0.64g/cm3；木材的强度和各项力学性能一般。
用途：用于高级家具和细木工制品、装饰材料、高级地板、乐器、雕刻制品等。
2.2 学名：缅茄Afzelia africana Smith.
商品名：道塞Doussie帕泡PapoKukpalik（加纳）Apa、Alinga（尼日
利亚）； Azodau、 Lingue（科特迪瓦）；Chamfuta、Mussacossa（莫桑比克）； Mbembakofi、 Mkora（坦桑比克）；M’bangaLingue（喀麦隆）；Afzelia（利比里亚）； Bolenug （扎伊尔）；Nkokongo（刚果）。木材材性：木材是散孔材；边材浅黄白色，宽度5cm,心材红褐色，常有斑点，与边材区别明显；生长轮略明显；木材具有光泽，无特殊气味；纹理混交错；结构细且均匀；耐磨性强，非常耐虫蛀；干缩小，生材至炉干干缩率弦向4.4%,径向3.0%木材硬而重,气干密度约0.83g/cm3；木材较稳定，强度高。
用途：用于一些重要的建筑工程，特别是港口码头建筑物和桥梁建设；带有漂亮花纹图案的可用于制造家具和装饰板。
2.3 学名：刚果铁木、奥特山榄 Aurranglla congoensis A.Chev.
商品名：木库轮古Mukulungu莫比Moabi：Djave（尼日利亚）； Elang、Elanzok（喀麦隆）；Mfua（刚果）；Kungulu（安哥拉）；Kabulungu、Kondo-fino（扎伊尔）。
木材材性：木材是散孔材；边材浅灰色；宽度2-3cm；心材浅褐色或红褐色，通常与边材区别明显；生长轮不明显；木材光泽弱，无特殊气味；木材的耐磨性强，抗酸和抗蚁性强，纹理直略交错，结构细；木材干缩大，生材到炉干时的干缩率弦向7.4%,径向5.8%木材硬且得,气干0.88-0.99g/cm3韧性大，强度、弯曲强度和载荷冲击强度均很高。
用途：用于重型建筑、桥梁、枕木、电杆等。
2.4 学名 白梨柴龙树Apodytes dimidiana E.Mey
商品名：穆冈犹讷 Mugonyone；White paer；Pearwood。
俗称：瑞士梨木
木材材性：边材和心材界限不明显，生材颜色由浅白到黄褐色，且带有粉色色调，在空气的作用下变成浅灰褐色；木材纹理直，结构细而均匀；较耐磨；材质硬且较重，气干密度约0.72g/cm3；各项力学指标中等，载荷冲击强度较低。
用途：未作防腐处理时，忌用其制作室内用具。
2.5 学名：特氏古夷黄木 Guibourtia tessmanii J.Leonard
商品名：卜宾佳Bubinga；Essingang（喀麦隆）；Kevazingo（加蓬）；Waka（扎伊尔）；Akume（美国）；0veng（赤道几内亚）。
俗称：非洲花梨。
木材材性；木材是散孔材；边材奶白色，宽度为5-7cm，心材红褐色，常具深色条纹，心边材区别明显；生长轮略明显；木材具有光泽，无特殊气味；结构细且均匀，纹理址至略交错；干缩大，生材至炉干干缩率径向5.3%,弦向7.8%木材耐腐,但边材常有菌、虫危害；木材硬且重，平均气干密度约0.89-0.91g/cm3；木材的强度和各项力学性能均较高，尤其是木材的横纹抗拉强度。
用途：由木材本身的条纹和交错纹理共同构成特殊的美丽花纹，因而具有较高的装饰价值。
2.6 学名：筒状非洲楝 Entandrophragma cylindrium Sprague。
商品名：沙比利Aapele；Sapelewood、Ubilesan（尼日利亚）；Sapelli（喀麦隆）；Aboudikro（象牙海岸）；Assi、Dilolo（加蓬）；Libuyu、Bobwe（扎伊尔）；Pendwa（加纳）；Muyoveu（乌干达）。
木材材性 ：木材是散孔材；边材浅黄色；宽度7-10cm，心材新切面是粉红色，时间长变红褐色，心边材区别明显；生长轮不明显；木材具有光泽，新切面有雪松气味；木材纹理交错，径切面有黑色条状花纹或梅花状花纹；结构细且均匀；木材干缩大，径向4.6%；弦向7.4%；木材较耐腐，但边材易受粉蠹虫危害；木材较硬，重量中等，气干密度约0.67 g/cm3；木材的强度和各项力学指标较高。
用途：用于高级装饰材料、高级家具和地板、高级细木工制品。
2.7 学名：猴子果Tieghemella heckelii pierre。
商品名：马扣热Makore；Aganokwa （尼日利亚）；Baku、Abako、Edumo（加纳）；Makorou、Dumori（科特迪瓦）；Doukd、Okola（加蓬）。
木材材性：木材散孔材，边材色浅，宽度5-6cm，心材红褐色，心边材区别不明显；生长轮不明显；木材光泽强，无特殊气味；纹理直，部分具有交错纹理，结构细且均匀；木材的干缩甚大，稳定性好；材质硬且重，气干密度0.62-0.72 g/cm3；木材的耐久性极强，能抗白蚁，偶然会出现蓝变；木材的韧性，强度和各项力学性能强。
用途：建筑材料、地板、家具、细木工制口、精密仪器、室内装修、雕刻工艺口、玩具等。
2.8 学名：大美木豆Pericopsis elata Van Meeuwen。
商品名：阿夫莫西亚 Afrormosia；Assameal（法国、象牙海岸）；Ejen（喀麦隆）；Kokrodua、Awawai（加纳）；Obang（加蓬）；Ole、Bahala、Mohole（扎伊尔、荷兰）；Ayin（尼日利亚）。
俗称：非洲柚木、红豆柚。
木材材性：木材是散孔材；边材窄，宽度1.5-2.5cm,心材黄褐色至深褐色,与边材区别明显生长轮不明显木材具有光泽,无特殊气味纹理略斜至交错,结构甚细且均匀木材耐久性高,不易腐朽和受虫害干缩较大,生材到炉干时径向干缩率3.0%,弦向干缩6.4%木材略重,气干密度0.70-0.86g/cm3木材的强度和各项力学指标较高。
用途：该木材可以替代柚木制造需要强度高、稳定性强、抗虫蛀的木制品；不要同深色金属做在一起，以防金属腐蚀使木材变色。
2.9 学名：罂粟尼索桐Nesogordonia papaverifera R.
商品名：丹它Danta；0voue（喀麦隆）；欧图图0tutu（尼日利亚）；Kotibe（象牙海岸）；Olborbora（加蓬）；Taanya（扎伊尔）；Naouga（安哥拉）；Epro（加纳）。
木材材性：木材是散孔材；边材浅褐色，宽度5-7cm，心材红褐色，心边格区别明显；生长轮不明显；木材具光泽、无特殊气味；木材结构甚细且均匀，纹理交错，在每项切面上呈斑点形图案；较耐腐，能抗白蚁危害，但易受海生钻木动物侵蚀；木材干缩很大；材质硬且重，气干密度076-0.80G/cm3木材有韧性,强度和其他力学性能指标均很高，但载荷冲击强度较低。
用途：建筑耐久构件、室内外装修、家具、地板、食品、包装箱、工具柄、雕刻工艺品等。
2.10 学名：西非香脂树Copaifera salilounda Heck。
商品名：埃蒂Etomoe（利比里亚、科行迪瓦）；Olumi、Anzem、Andem-Evine（加蓬）；Ohwendua、Entedua（加纳）；Buini、Gumcoal（塞拉利昂）；Ovblaleke（尼日利亚）；Bofelele（扎伊尔）。
木材材性：木材散孔材；边材色浅，宽度6-10cm，心材红褐色，常有深色条纹，心边材区别明显；生长轮略明显；木材具有光泽，无特殊气味；干缩大，生材到炉干干缩率径向为4.5%,弦向为7.5%木材纹理直至略交错,结构甚细而均匀木材耐腐性能和抗蚁性能一般木材重量较重,气干密度约07.8g/cm3强度和各英力学性能良好。
用途：建筑构件、室内装修等。
2.11 学名：高贵绿柄桑Chlorophora regiaA.Chev
商品名：奥贵绿柄桑Odum（加纳、象牙海岸）；埃若科Irolo、Semli（塞拉利昂、利比里亚）Rokko、Oroko、（尼日利亚）；Abang、Mandji（喀麦隆）；Kambala（扎伊尔）；Mereira（安哥拉）；Mvule（东非）。
俗称：非洲黄金木（不很确定）
木材材性；木材是散孔格；边材黄白色，宽度5cm ，心材新切面是黄色或浅褐色，久露空气中成为金黄褐色，心边材略有区别；生长轮不明显；木材具有光泽，无特殊气味；耐腐性好，不宜受小蠹虫危害；木材干缩小至中，生材至炉干干缩率径向2.1-4.0%；弦向3.6-6.5%；木材纹理斜或交错,结构略细且略均匀；在由机械损伤而造成的木材裂痕和沟槽中，有碳酸钙沉淀物（称为：石头）；木材重量中等，平均气干密度约0.66g/cm3；强度和各项力池指标较好。
用途：家具、地板、细木工制品、胶合板、门、窗、造、船、枕木、雕刻工艺品等。
2.12 学名：斯图崖豆木Millettia stuhlmannii Taub
商品名：番加-番加Panga-panga；詹母贝尔Jambire；Partridgewood；
Mpande。
俗称：非洲鸡翅木。
木材材性：木材是散孔材；边材为浅黄煞费苦心，宽度为2.5-7.5cm；心材为巧克力且有深浅间隔的色带，或是深褐色且有白色的色带；心边材区别明显，生长轮明显；木材上有一种特殊的鹑鸡羽毛花纹；木材一般纹理直，结构中等且均匀，无特殊气味；很耐腐；具有天然的抗真菌腐化和白蚁的蛀蚀；有丰富的深色树胶；材质硬且重，气干密度0.80-0.91g/cm3木材高度耐磨，干缩小，强度中等。
用途：该木材用于重型建筑、重载地板、高档家具和装饰材料、钢琴、小提琴、胶合板、精密仪器、雕刻品等。
3 美洲木材（12种）
欢迎分享，转载请注明来源：夏雨云
微信扫一扫
支付宝扫一扫
评论列表（0条）